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

Diff of /code/trunk/ChangeLog

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

revision 11 by nigel, Sat Feb 24 21:38:17 2007 UTC revision 775 by ph10, Thu Dec 1 10:35:30 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". The
60        maximum number of forward references (including repeats) was limited by the
61        internal workspace, and dependent on the LINK_SIZE. The code has been
62        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
63        the default depends on LINK_SIZE. There is a new upper limit (for safety)
64        of around 200,000 forward references. While doing this, I also speeded up
65        the filling in of repeated forward references.
66    
67    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
68        incorrectly expecting the subject to contain another "a" after the start.
69    
70    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
71        in the match, the SKIP should be ignored. This was not happening; instead
72        the SKIP was being treated as NOMATCH. For patterns such as
73        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
74        tested.
75    
76    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
77        now much more compatible with Perl, in particular in cases where the result
78        is a non-match for a non-anchored pattern. For example, if
79        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
80        "m", where previously it did not return a name. A side effect of this
81        change is that for partial matches, the last encountered mark name is
82        returned, as for non matches. A number of tests that were previously not
83        Perl-compatible have been moved into the Perl-compatible test files. The
84        refactoring has had the pleasing side effect of removing one argument from
85        the match() function, thus reducing its stack requirements.
86    
87    19. If the /S+ option was used in pcretest to study a pattern using JIT,
88        subsequent uses of /S (without +) incorrectly behaved like /S+.
89    
90    
91    Version 8.20 21-Oct-2011
92    ------------------------
93    
94    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
95        a POSIX class. After further experiments with Perl, which convinced me that
96        Perl has bugs and confusions, a closing square bracket is no longer allowed
97        in a POSIX name. This bug also affected patterns with classes that started
98        with full stops.
99    
100    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
101        captured substring, but while checking the failing first alternative,
102        substring 1 is temporarily captured. If the output vector supplied to
103        pcre_exec() was not big enough for this capture, the yield of the function
104        was still zero ("insufficient space for captured substrings"). This cannot
105        be totally fixed without adding another stack variable, which seems a lot
106        of expense for a edge case. However, I have improved the situation in cases
107        such as /(a)(b)x|abc/ matched against "abc", where the return code
108        indicates that fewer than the maximum number of slots in the ovector have
109        been set.
110    
111    3.  Related to (2) above: when there are more back references in a pattern than
112        slots in the output vector, pcre_exec() uses temporary memory during
113        matching, and copies in the captures as far as possible afterwards. It was
114        using the entire output vector, but this conflicts with the specification
115        that only 2/3 is used for passing back captured substrings. Now it uses
116        only the first 2/3, for compatibility. This is, of course, another edge
117        case.
118    
119    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
120        main code base, and can be used by building with --enable-jit. When this is
121        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
122        runtime --no-jit option is given.
123    
124    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
125        ovector, the return from the function was zero, implying that there were
126        other matches that did not fit. The correct "exactly full" value is now
127        returned.
128    
129    6.  If a subpattern that was called recursively or as a subroutine contained
130        (*PRUNE) or any other control that caused it to give a non-standard return,
131        invalid errors such as "Error -26 (nested recursion at the same subject
132        position)" or even infinite loops could occur.
133    
134    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
135        computing the minimum length on reaching *ACCEPT, and so ended up with the
136        wrong value of 1 rather than 0. Further investigation indicates that
137        computing a minimum subject length in the presence of *ACCEPT is difficult
138        (think back references, subroutine calls), and so I have changed the code
139        so that no minimum is registered for a pattern that contains *ACCEPT.
140    
141    8.  If (*THEN) was present in the first (true) branch of a conditional group,
142        it was not handled as intended. [But see 16 below.]
143    
144    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
145        Sheri Pierce.
146    
147    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
148        the first byte in a match must be "a".
149    
150    11. Change 17 for 8.13 increased the recursion depth for patterns like
151        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
152        pattern contains any instances of (*THEN). If it does not, the old
153        optimizations are restored. It would be nice to do this on a per-group
154        basis, but at the moment that is not feasible.
155    
156    12. In some environments, the output of pcretest -C is CRLF terminated. This
157        broke RunTest's code that checks for the link size. A single white space
158        character after the value is now allowed for.
159    
160    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
161        For "fr", it uses the Windows-specific input and output files.
162    
163    14. If (*THEN) appeared in a group that was called recursively or as a
164        subroutine, it did not work as intended. [But see next item.]
165    
166    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
167        pattern fragments (but not containing any | characters). If A and B are
168        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
169        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
170        D. In other words, Perl considers parentheses that do not contain any |
171        characters to be part of a surrounding alternative, whereas PCRE was
172        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
173        differently. PCRE now behaves in the same way as Perl, except in the case
174        of subroutine/recursion calls such as (?1) which have in any case always
175        been different (but PCRE had them first :-).
176    
177    16. Related to 15 above: Perl does not treat the | in a conditional group as
178        creating alternatives. Such a group is treated in the same way as an
179        ordinary group without any | characters when processing (*THEN). PCRE has
180        been changed to match Perl's behaviour.
181    
182    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
183        RunGrepTest script failed.
184    
185    18. Change 22 for version 13 caused atomic groups to use more stack. This is
186        inevitable for groups that contain captures, but it can lead to a lot of
187        stack use in large patterns. The old behaviour has been restored for atomic
188        groups that do not contain any capturing parentheses.
189    
190    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
191        suppress the check for a minimum subject length at run time. (If it was
192        given to pcre_exec() or pcre_dfa_exec() it did work.)
193    
194    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
195        fail to work when decoding hex characters in data strings in EBCDIC
196        environments.
197    
198    21. It appears that in at least one Mac OS environment, the isxdigit() function
199        is implemented as a macro that evaluates to its argument more than once,
200        contravening the C 90 Standard (I haven't checked a later standard). There
201        was an instance in pcretest which caused it to go wrong when processing
202        \x{...} escapes in subject strings. The has been rewritten to avoid using
203        things like p++ in the argument of isxdigit().
204    
205    
206    Version 8.13 16-Aug-2011
207    ------------------------
208    
209    1.  The Unicode data tables have been updated to Unicode 6.0.0.
210    
211    2.  Two minor typos in pcre_internal.h have been fixed.
212    
213    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
214        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
215        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
216    
217    4.  There were a number of related bugs in the code for matching backrefences
218        caselessly in UTF-8 mode when codes for the characters concerned were
219        different numbers of bytes. For example, U+023A and U+2C65 are an upper
220        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
221        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
222        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
223        2-byte code at the end of the subject (it thought there wasn't enough data
224        left).
225    
226    5.  Comprehensive information about what went wrong is now returned by
227        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
228        as the output vector has at least 2 elements. The offset of the start of
229        the failing character and a reason code are placed in the vector.
230    
231    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
232        now returned is for the first byte of the failing character, instead of the
233        last byte inspected. This is an incompatible change, but I hope it is small
234        enough not to be a problem. It makes the returned offset consistent with
235        pcre_exec() and pcre_dfa_exec().
236    
237    7.  pcretest now gives a text phrase as well as the error number when
238        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
239        failure, the offset and reason code are output.
240    
241    8.  When \R was used with a maximizing quantifier it failed to skip backwards
242        over a \r\n pair if the subsequent match failed. Instead, it just skipped
243        back over a single character (\n). This seems wrong (because it treated the
244        two characters as a single entity when going forwards), conflicts with the
245        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
246        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
247        has been changed.
248    
249    9.  Some internal refactoring has changed the processing so that the handling
250        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
251        time (the PCRE_DOTALL option was changed this way some time ago: version
252        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
253        which was always a bit of a fudge. It also means that there is one less
254        argument for the match() function, which reduces its stack requirements
255        slightly. This change also fixes an incompatibility with Perl: the pattern
256        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
257    
258    10. More internal refactoring has drastically reduced the number of recursive
259        calls to match() for possessively repeated groups such as (abc)++ when
260        using pcre_exec().
261    
262    11. While implementing 10, a number of bugs in the handling of groups were
263        discovered and fixed:
264    
265        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
266        (a|)*(?1) gave a compile-time internal error.
267        ((a|)+)+  did not notice that the outer group could match an empty string.
268        (^a|^)+   was not marked as anchored.
269        (.*a|.*)+ was not marked as matching at start or after a newline.
270    
271    12. Yet more internal refactoring has removed another argument from the match()
272        function. Special calls to this function are now indicated by setting a
273        value in a variable in the "match data" data block.
274    
275    13. Be more explicit in pcre_study() instead of relying on "default" for
276        opcodes that mean there is no starting character; this means that when new
277        ones are added and accidentally left out of pcre_study(), testing should
278        pick them up.
279    
280    14. The -s option of pcretest has been documented for ages as being an old
281        synonym of -m (show memory usage). I have changed it to mean "force study
282        for every regex", that is, assume /S for every regex. This is similar to -i
283        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
284        using it. It makes it easier to run collections of tests with and without
285        study enabled, and thereby test pcre_study() more easily. All the standard
286        tests are now run with and without -s (but some patterns can be marked as
287        "never study" - see 20 below).
288    
289    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
290        restoration of the capturing data to the outer values was not happening
291        correctly.
292    
293    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
294        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
295        pattern had matched an empty string, and so incorrectly returned a no
296        match.
297    
298    17. There was optimizing code for the last branch of non-capturing parentheses,
299        and also for the obeyed branch of a conditional subexpression, which used
300        tail recursion to cut down on stack usage. Unfortunately, now that there is
301        the possibility of (*THEN) occurring in these branches, tail recursion is
302        no longer possible because the return has to be checked for (*THEN). These
303        two optimizations have therefore been removed. [But see 8.20/11 above.]
304    
305    18. If a pattern containing \R was studied, it was assumed that \R always
306        matched two bytes, thus causing the minimum subject length to be
307        incorrectly computed because \R can also match just one byte.
308    
309    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
310        was incorrectly computed.
311    
312    20. If /S is present twice on a test pattern in pcretest input, it now
313        *disables* studying, thereby overriding the use of -s on the command line
314        (see 14 above). This is necessary for one or two tests to keep the output
315        identical in both cases.
316    
317    21. When (*ACCEPT) was used in an assertion that matched an empty string and
318        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
319    
320    22. When an atomic group that contained a capturing parenthesis was
321        successfully matched, but the branch in which it appeared failed, the
322        capturing was not being forgotten if a higher numbered group was later
323        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
324        group 1 to "a", when in fact it should be unset. This applied to multi-
325        branched capturing and non-capturing groups, repeated or not, and also to
326        positive assertions (capturing in negative assertions does not happen
327        in PCRE) and also to nested atomic groups.
328    
329    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
330        subject after a captured substring, to make it easier to tell which of a
331        number of identical substrings has been captured.
332    
333    24. The way atomic groups are processed by pcre_exec() has been changed so that
334        if they are repeated, backtracking one repetition now resets captured
335        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
336        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
337        "aaa". Previously, it would have been "a". As part of this code
338        refactoring, the way recursive calls are handled has also been changed.
339    
340    25. If an assertion condition captured any substrings, they were not passed
341        back unless some other capturing happened later. For example, if
342        (?(?=(a))a) was matched against "a", no capturing was returned.
343    
344    26. When studying a pattern that contained subroutine calls or assertions,
345        the code for finding the minimum length of a possible match was handling
346        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
347        group 1 called group 2 while simultaneously a separate group 2 called group
348        1). A stack overflow occurred in this case. I have fixed this by limiting
349        the recursion depth to 10.
350    
351    27. Updated RunTest.bat in the distribution to the version supplied by Tom
352        Fortmann. This supports explicit test numbers on the command line, and has
353        argument validation and error reporting.
354    
355    28. An instance of \X with an unlimited repeat could fail if at any point the
356        first character it looked at was a mark character.
357    
358    29. Some minor code refactoring concerning Unicode properties and scripts
359        should reduce the stack requirement of match() slightly.
360    
361    30. Added the '=' option to pcretest to check the setting of unused capturing
362        slots at the end of the pattern, which are documented as being -1, but are
363        not included in the return count.
364    
365    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
366        compiled something random. Now it gives a compile-time error (as does
367        Perl).
368    
369    32. A *MARK encountered during the processing of a positive assertion is now
370        recorded and passed back (compatible with Perl).
371    
372    33. If --only-matching or --colour was set on a pcregrep call whose pattern
373        had alternative anchored branches, the search for a second match in a line
374        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
375        matched the line "0102" twice. The same bug affected patterns that started
376        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
377        twice.
378    
379    34. Previously, PCRE did not allow quantification of assertions. However, Perl
380        does, and because of capturing effects, quantifying parenthesized
381        assertions may at times be useful. Quantifiers are now allowed for
382        parenthesized assertions.
383    
384    35. A minor code tidy in pcre_compile() when checking options for \R usage.
385    
386    36. \g was being checked for fancy things in a character class, when it should
387        just be a literal "g".
388    
389    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
390        appearance of a nested POSIX class supersedes an apparent external class.
391        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
392        unescaped square brackets may also appear as part of class names. For
393        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
394        more like Perl. (But see 8.20/1 above.)
395    
396    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
397        was because it thought it was \N{name}, which is not supported).
398    
399    39. Add minix to OS list not supporting the -S option in pcretest.
400    
401    40. PCRE tries to detect cases of infinite recursion at compile time, but it
402        cannot analyze patterns in sufficient detail to catch mutual recursions
403        such as ((?1))((?2)). There is now a runtime test that gives an error if a
404        subgroup is called recursively as a subpattern for a second time at the
405        same position in the subject string. In previous releases this might have
406        been caught by the recursion limit, or it might have run out of stack.
407    
408    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
409        happen only once. PCRE was, however incorrectly giving a compile time error
410        "recursive call could loop indefinitely" because it cannot analyze the
411        pattern in sufficient detail. The compile time test no longer happens when
412        PCRE is compiling a conditional subpattern, but actual runaway loops are
413        now caught at runtime (see 40 above).
414    
415    42. It seems that Perl allows any characters other than a closing parenthesis
416        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
417        has been changed to be the same.
418    
419    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
420        as not to get warnings when autogen.sh is called. Also changed
421        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
422    
423    44. To help people who use pcregrep to scan files containing exceedingly long
424        lines, the following changes have been made:
425    
426        (a) The default value of the buffer size parameter has been increased from
427            8K to 20K. (The actual buffer used is three times this size.)
428    
429        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
430            PCRE is built.
431    
432        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
433            to be set at run time.
434    
435        (d) Numerical values in pcregrep options can be followed by K or M, for
436            example --buffer-size=50K.
437    
438        (e) If a line being scanned overflows pcregrep's buffer, an error is now
439            given and the return code is set to 2.
440    
441    45. Add a pointer to the latest mark to the callout data block.
442    
443    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
444        partial match of an empty string instead of no match. This was specific to
445        the use of ".".
446    
447    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
448        complete match instead of a partial match. This bug was dependent on both
449        the PCRE_UTF8 and PCRE_DOTALL options being set.
450    
451    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
452        starting byte set, because \b was not being ignored.
453    
454    
455    Version 8.12 15-Jan-2011
456    ------------------------
457    
458    1.  Fixed some typos in the markup of the man pages, and wrote a script that
459        checks for such things as part of the documentation building process.
460    
461    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
462        --match-limit and --recursion-limit options (added for 8.11). In
463        particular, this made one of the standard tests fail. (The integer value
464        went into the wrong half of a long int.)
465    
466    3.  If the --colour option was given to pcregrep with -v (invert match), it
467        did strange things, either producing crazy output, or crashing. It should,
468        of course, ignore a request for colour when reporting lines that do not
469        match.
470    
471    4.  Another pcregrep bug caused similar problems if --colour was specified with
472        -M (multiline) and the pattern match finished with a line ending.
473    
474    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
475        matched in multiline mode, the following line was shown as part of the
476        match. This seems wrong, so I have changed it.
477    
478    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
479        the check for further matches in the same line (so they could be coloured)
480        to overrun the end of the current line. If another match was found, it was
481        incorrectly shown (and then shown again when found in the next line).
482    
483    7.  If pcregrep was compiled under Windows, there was a reference to the
484        function pcregrep_exit() before it was defined. I am assuming this was
485        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
486        reported by a user. I've moved the definition above the reference.
487    
488    
489    Version 8.11 10-Dec-2010
490    ------------------------
491    
492    1.  (*THEN) was not working properly if there were untried alternatives prior
493        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
494        backtracked to try for "b" instead of moving to the next alternative branch
495        at the same level (in this case, to look for "c"). The Perl documentation
496        is clear that when (*THEN) is backtracked onto, it goes to the "next
497        alternative in the innermost enclosing group".
498    
499    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
500        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
501        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
502        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
503        (*THEN).
504    
505    3.  If \s appeared in a character class, it removed the VT character from
506        the class, even if it had been included by some previous item, for example
507        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
508        of \s, but is part of the POSIX "space" class.)
509    
510    4.  A partial match never returns an empty string (because you can always
511        match an empty string at the end of the subject); however the checking for
512        an empty string was starting at the "start of match" point. This has been
513        changed to the "earliest inspected character" point, because the returned
514        data for a partial match starts at this character. This means that, for
515        example, /(?<=abc)def/ gives a partial match for the subject "abc"
516        (previously it gave "no match").
517    
518    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
519        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
520        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
521        has an implication that the given string is incomplete (because a partial
522        match is preferred over a full match). For this reason, these items now
523        give a partial match in this situation. [Aside: previously, the one case
524        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
525        match rather than a full match, which was wrong by the old rules, but is
526        now correct.]
527    
528    6.  There was a bug in the handling of #-introduced comments, recognized when
529        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
530        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
531        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
532        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
533        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
534        places in pcre_compile().
535    
536    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
537        comments when looking ahead for named forward references to subpatterns,
538        the only newline sequence it recognized was NL. It now handles newlines
539        according to the set newline convention.
540    
541    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
542        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
543        cater for a lack of strerror(). These oversights have been fixed.
544    
545    9.  Added --match-limit and --recursion-limit to pcregrep.
546    
547    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
548    
549    11. When the -o option was used, pcregrep was setting a return code of 1, even
550        when matches were found, and --line-buffered was not being honoured.
551    
552    12. Added an optional parentheses number to the -o and --only-matching options
553        of pcregrep.
554    
555    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
556        can match an empty string. The code to do it in pcretest and pcredemo
557        needed fixing:
558    
559        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
560            only one byte after an empty string match just before CRLF (this case
561            just got forgotten; "any" and "anycrlf" were OK).
562    
563        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
564            mode when an empty string match preceded an ASCII character followed by
565            a non-ASCII character. (The code for advancing by one character rather
566            than one byte was nonsense.)
567    
568        (c) The pcredemo.c sample program did not have any code at all to handle
569            the cases when CRLF is a valid newline sequence.
570    
571    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
572        as a starting offset was within the subject string. There is now a new
573        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
574        negative or greater than the length of the string. In order to test this,
575        pcretest is extended to allow the setting of negative starting offsets.
576    
577    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
578        starting offset points to the beginning of a UTF-8 character was
579        unnecessarily clumsy. I tidied it up.
580    
581    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
582        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
583    
584    17. Nobody had reported that the --include_dir option, which was added in
585        release 7.7 should have been called --include-dir (hyphen, not underscore)
586        for compatibility with GNU grep. I have changed it to --include-dir, but
587        left --include_dir as an undocumented synonym, and the same for
588        --exclude-dir, though that is not available in GNU grep, at least as of
589        release 2.5.4.
590    
591    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
592        characters from a string of bytes) have been redefined so as not to use
593        loops, in order to improve performance in some environments. At the same
594        time, I abstracted some of the common code into auxiliary macros to save
595        repetition (this should not affect the compiled code).
596    
597    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
598        compile-time error is now given if \c is not followed by an ASCII
599        character, that is, a byte less than 128. (In EBCDIC mode, the code is
600        different, and any byte value is allowed.)
601    
602    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
603        START_OPTIMIZE option, which is now allowed at compile time - but just
604        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
605        to pcregrep and other applications that have no direct access to PCRE
606        options. The new /Y option in pcretest sets this option when calling
607        pcre_compile().
608    
609    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
610        back references. Groups containing recursive back references were forced to
611        be atomic by that change, but in the case of named groups, the amount of
612        memory required was incorrectly computed, leading to "Failed: internal
613        error: code overflow". This has been fixed.
614    
615    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
616        pcretest.c, to avoid build problems in some Borland environments.
617    
618    
619    Version 8.10 25-Jun-2010
620    ------------------------
621    
622    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
623        THEN.
624    
625    2.  (*ACCEPT) was not working when inside an atomic group.
626    
627    3.  Inside a character class, \B is treated as a literal by default, but
628        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
629        causes the error). The code is unchanged, but I tidied the documentation.
630    
631    4.  Inside a character class, PCRE always treated \R and \X as literals,
632        whereas Perl faults them if its -w option is set. I have changed PCRE so
633        that it faults them when PCRE_EXTRA is set.
634    
635    5.  Added support for \N, which always matches any character other than
636        newline. (It is the same as "." when PCRE_DOTALL is not set.)
637    
638    6.  When compiling pcregrep with newer versions of gcc which may have
639        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
640        declared with attribute warn_unused_result" were given. Just casting the
641        result to (void) does not stop the warnings; a more elaborate fudge is
642        needed. I've used a macro to implement this.
643    
644    7.  Minor change to pcretest.c to avoid a compiler warning.
645    
646    8.  Added four artifical Unicode properties to help with an option to make
647        \s etc use properties (see next item). The new properties are: Xan
648        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
649    
650    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
651        use Unicode properties. (*UCP) at the start of a pattern can be used to set
652        this option. Modified pcretest to add /W to test this facility. Added
653        REG_UCP to make it available via the POSIX interface.
654    
655    10. Added --line-buffered to pcregrep.
656    
657    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
658        studied, and the match started with a letter with a code point greater than
659        127 whose first byte was different to the first byte of the other case of
660        the letter, the other case of this starting letter was not recognized
661        (#976).
662    
663    12. If a pattern that was studied started with a repeated Unicode property
664        test, for example, \p{Nd}+, there was the theoretical possibility of
665        setting up an incorrect bitmap of starting bytes, but fortunately it could
666        not have actually happened in practice until change 8 above was made (it
667        added property types that matched character-matching opcodes).
668    
669    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
670        possible starting bytes for non-anchored patterns.
671    
672    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
673        \R, and also a number of cases that involve Unicode properties, both
674        explicit and implicit when PCRE_UCP is set.
675    
676    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
677        input, it could crash or give wrong results if characters with values
678        greater than 0xc0 were present in the subject string. (Detail: it assumed
679        UTF-8 input when processing these items.)
680    
681    16. Added a lot of (int) casts to avoid compiler warnings in systems where
682        size_t is 64-bit (#991).
683    
684    17. Added a check for running out of memory when PCRE is compiled with
685        --disable-stack-for-recursion (#990).
686    
687    18. If the last data line in a file for pcretest does not have a newline on
688        the end, a newline was missing in the output.
689    
690    19. The default pcre_chartables.c file recognizes only ASCII characters (values
691        less than 128) in its various bitmaps. However, there is a facility for
692        generating tables according to the current locale when PCRE is compiled. It
693        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
694        characters, are recognized by isspace() and therefore were getting set in
695        these tables, and indeed these tables seem to approximate to ISO 8859. This
696        caused a problem in UTF-8 mode when pcre_study() was used to create a list
697        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
698        which of course cannot start UTF-8 characters. I have changed the code so
699        that only real ASCII characters (less than 128) and the correct starting
700        bytes for UTF-8 encodings are set for characters greater than 127 when in
701        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
702        altogether.)
703    
704    20. Added the /T option to pcretest so as to be able to run tests with non-
705        standard character tables, thus making it possible to include the tests
706        used for 19 above in the standard set of tests.
707    
708    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
709        reference to a subpattern the other side of a comment that contains an
710        opening parenthesis caused either an internal compiling error, or a
711        reference to the wrong subpattern.
712    
713    
714    Version 8.02 19-Mar-2010
715    ------------------------
716    
717    1.  The Unicode data tables have been updated to Unicode 5.2.0.
718    
719    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
720        configured.
721    
722    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
723        original author of that file, following a query about its status.
724    
725    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
726        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
727    
728    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
729        quantifier applied to a forward-referencing subroutine call, could compile
730        incorrect code or give the error "internal error: previously-checked
731        referenced subpattern not found".
732    
733    6.  Both MS Visual Studio and Symbian OS have problems with initializing
734        variables to point to external functions. For these systems, therefore,
735        pcre_malloc etc. are now initialized to local functions that call the
736        relevant global functions.
737    
738    7.  There were two entries missing in the vectors called coptable and poptable
739        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
740        I've fixed the data, and added a kludgy way of testing at compile time that
741        the lengths are correct (equal to the number of opcodes).
742    
743    8.  Following on from 7, I added a similar kludge to check the length of the
744        eint vector in pcreposix.c.
745    
746    9.  Error texts for pcre_compile() are held as one long string to avoid too
747        much relocation at load time. To find a text, the string is searched,
748        counting zeros. There was no check for running off the end of the string,
749        which could happen if a new error number was added without updating the
750        string.
751    
752    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
753    
754    11. \K was not working if it appeared in an atomic group or in a group that
755        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
756        \K is "not well defined" if used in an assertion. PCRE now accepts it if
757        the assertion is positive, but not if it is negative.
758    
759    12. Change 11 fortuitously reduced the size of the stack frame used in the
760        "match()" function of pcre_exec.c by one pointer. Forthcoming
761        implementation of support for (*MARK) will need an extra pointer on the
762        stack; I have reserved it now, so that the stack frame size does not
763        decrease.
764    
765    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
766        item in branch that calls a recursion is a subroutine call - as in the
767        second branch in the above example - was incorrectly given the compile-
768        time error "recursive call could loop indefinitely" because pcre_compile()
769        was not correctly checking the subroutine for matching a non-empty string.
770    
771    14. The checks for overrunning compiling workspace could trigger after an
772        overrun had occurred. This is a "should never occur" error, but it can be
773        triggered by pathological patterns such as hundreds of nested parentheses.
774        The checks now trigger 100 bytes before the end of the workspace.
775    
776    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
777    
778    
779    Version 8.01 19-Jan-2010
780    ------------------------
781    
782    1.  If a pattern contained a conditional subpattern with only one branch (in
783        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
784        computed the wrong minimum data length (which is of course zero for such
785        subpatterns). This could cause incorrect "no match" results.
786    
787    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
788        the pattern is reset in the first branch, pcre_compile() failed with
789        "internal error: code overflow at offset...". This happened only when
790        the reset was to the original external option setting. (An optimization
791        abstracts leading options settings into an external setting, which was the
792        cause of this.)
793    
794    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
795        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
796        assertion pattern did not match (meaning that the assertion was true), it
797        was incorrectly treated as false if the SKIP had been reached during the
798        matching. This also applied to assertions used as conditions.
799    
800    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
801        assertion subpattern, including such a pattern used as a condition,
802        unpredictable results occurred, instead of the error return
803        PCRE_ERROR_DFA_UITEM.
804    
805    5.  The C++ GlobalReplace function was not working like Perl for the special
806        situation when an empty string is matched. It now does the fancy magic
807        stuff that is necessary.
808    
809    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
810        removed. (These were left over from very, very early versions of PCRE.)
811    
812    7.  Some cosmetic changes to the code to make life easier when compiling it
813        as part of something else:
814    
815        (a) Change DEBUG to PCRE_DEBUG.
816    
817        (b) In pcre_compile(), rename the member of the "branch_chain" structure
818            called "current" as "current_branch", to prevent a collision with the
819            Linux macro when compiled as a kernel module.
820    
821        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
822            prevent a collision with the Linux macro when compiled as a kernel
823            module.
824    
825    8.  In pcre_compile() there are some checks for integer overflows that used to
826        cast potentially large values to (double). This has been changed to that
827        when building, a check for int64_t is made, and if it is found, it is used
828        instead, thus avoiding the use of floating point arithmetic. (There is no
829        other use of FP in PCRE.) If int64_t is not found, the fallback is to
830        double.
831    
832    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
833        2005 (difference between two addresses compared to an unsigned value).
834    
835    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
836        custom one, because of the following reported problem in Windows:
837    
838          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
839              under Win32.
840          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
841              therefore missing the function definition.
842          - The compiler thus generates a "C" signature for the test function.
843          - The linker fails to find the "C" function.
844          - PCRE fails to configure if asked to do so against libbz2.
845    
846    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
847        messages were output:
848    
849          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
850          rerunning libtoolize, to keep the correct libtool macros in-tree.
851          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
852    
853        I have done both of these things.
854    
855    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
856        most of the time, it *can* run out if it is given a pattern that contains a
857        runaway infinite recursion. I updated the discussion in the pcrestack man
858        page.
859    
860    13. Now that we have gone to the x.xx style of version numbers, the minor
861        version may start with zero. Using 08 or 09 is a bad idea because users
862        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
863        interpreted as invalid octal numbers. I've updated the previous comment in
864        configure.ac, and also added a check that gives an error if 08 or 09 are
865        used.
866    
867    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
868        causing partial matching to fail when the end of the subject matched \W
869        in a UTF-8 pattern where \W was quantified with a minimum of 3.
870    
871    15. There were some discrepancies between the declarations in pcre_internal.h
872        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
873        their definitions. The declarations used "const uschar *" and the
874        definitions used USPTR. Even though USPTR is normally defined as "const
875        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
876        reported that: "This difference in casting confuses some C++ compilers, for
877        example, SunCC recognizes above declarations as different functions and
878        generates broken code for hbpcre." I have changed the declarations to use
879        USPTR.
880    
881    16. GNU libtool is named differently on some systems. The autogen.sh script now
882        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
883        (FreeBSD).
884    
885    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
886        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
887        comment: "Figure out how to create a longlong from a string: strtoll and
888        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
889        instance, but it only takes 2 args instead of 3!"
890    
891    18. A subtle bug concerned with back references has been fixed by a change of
892        specification, with a corresponding code fix. A pattern such as
893        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
894        refers, was giving matches when it shouldn't. For example, xa=xaaa would
895        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
896        same bug. Such groups have to be quantified to be useful, or contained
897        inside another quantified group. (If there's no repetition, the reference
898        can never match.) The problem arises because, having left the group and
899        moved on to the rest of the pattern, a later failure that backtracks into
900        the group uses the captured value from the final iteration of the group
901        rather than the correct earlier one. I have fixed this in PCRE by forcing
902        any group that contains a reference to itself to be an atomic group; that
903        is, there cannot be any backtracking into it once it has completed. This is
904        similar to recursive and subroutine calls.
905    
906    
907    Version 8.00 19-Oct-09
908    ----------------------
909    
910    1.  The table for translating pcre_compile() error codes into POSIX error codes
911        was out-of-date, and there was no check on the pcre_compile() error code
912        being within the table. This could lead to an OK return being given in
913        error.
914    
915    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
916        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
917        in a Windows environment.
918    
919    3.  The pcregrep --count option prints the count for each file even when it is
920        zero, as does GNU grep. However, pcregrep was also printing all files when
921        --files-with-matches was added. Now, when both options are given, it prints
922        counts only for those files that have at least one match. (GNU grep just
923        prints the file name in this circumstance, but including the count seems
924        more useful - otherwise, why use --count?) Also ensured that the
925        combination -clh just lists non-zero counts, with no names.
926    
927    4.  The long form of the pcregrep -F option was incorrectly implemented as
928        --fixed_strings instead of --fixed-strings. This is an incompatible change,
929        but it seems right to fix it, and I didn't think it was worth preserving
930        the old behaviour.
931    
932    5.  The command line items --regex=pattern and --regexp=pattern were not
933        recognized by pcregrep, which required --regex pattern or --regexp pattern
934        (with a space rather than an '='). The man page documented the '=' forms,
935        which are compatible with GNU grep; these now work.
936    
937    6.  No libpcreposix.pc file was created for pkg-config; there was just
938        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
939    
940    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
941        when UCP support is not needed, by modifying the Python script that
942        generates it from Unicode data files. This should not matter if the module
943        is correctly used as a library, but I received one complaint about 50K of
944        unwanted data. My guess is that the person linked everything into his
945        program rather than using a library. Anyway, it does no harm.
946    
947    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
948        was a minimum greater than 1 for a wide character in a possessive
949        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
950        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
951        character. Chaos in the form of incorrect output or a compiling loop could
952        result.
953    
954    9.  The restrictions on what a pattern can contain when partial matching is
955        requested for pcre_exec() have been removed. All patterns can now be
956        partially matched by this function. In addition, if there are at least two
957        slots in the offset vector, the offset of the earliest inspected character
958        for the match and the offset of the end of the subject are set in them when
959        PCRE_ERROR_PARTIAL is returned.
960    
961    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
962        synonymous with PCRE_PARTIAL, for backwards compatibility, and
963        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
964        and may be more useful for multi-segment matching.
965    
966    11. Partial matching with pcre_exec() is now more intuitive. A partial match
967        used to be given if ever the end of the subject was reached; now it is
968        given only if matching could not proceed because another character was
969        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
970        string "Z", which now yields "no match" instead of "partial match". In the
971        case of pcre_dfa_exec(), "no match" is given if every matching path for the
972        final character ended with (*FAIL).
973    
974    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
975        if the pattern had a "must contain" character that was already found in the
976        earlier partial match, unless partial matching was again requested. For
977        example, with the pattern /dog.(body)?/, the "must contain" character is
978        "g". If the first part-match was for the string "dog", restarting with
979        "sbody" failed. This bug has been fixed.
980    
981    13. The string returned by pcre_dfa_exec() after a partial match has been
982        changed so that it starts at the first inspected character rather than the
983        first character of the match. This makes a difference only if the pattern
984        starts with a lookbehind assertion or \b or \B (\K is not supported by
985        pcre_dfa_exec()). It's an incompatible change, but it makes the two
986        matching functions compatible, and I think it's the right thing to do.
987    
988    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
989        so that the demonstration program is easily available in environments where
990        PCRE has not been installed from source.
991    
992    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
993        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
994        library.
995    
996    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
997        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
998        is not the first non-POSIX option to be added. Clearly some people find
999        these options useful.
1000    
1001    17. If a caller to the POSIX matching function regexec() passes a non-zero
1002        value for nmatch with a NULL value for pmatch, the value of
1003        nmatch is forced to zero.
1004    
1005    18. RunGrepTest did not have a test for the availability of the -u option of
1006        the diff command, as RunTest does. It now checks in the same way as
1007        RunTest, and also checks for the -b option.
1008    
1009    19. If an odd number of negated classes containing just a single character
1010        interposed, within parentheses, between a forward reference to a named
1011        subpattern and the definition of the subpattern, compilation crashed with
1012        an internal error, complaining that it could not find the referenced
1013        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1014        [The bug was that it was starting one character too far in when skipping
1015        over the character class, thus treating the ] as data rather than
1016        terminating the class. This meant it could skip too much.]
1017    
1018    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1019        /g option in pcretest when the pattern contains \K, which makes it possible
1020        to have an empty string match not at the start, even when the pattern is
1021        anchored. Updated pcretest and pcredemo to use this option.
1022    
1023    21. If the maximum number of capturing subpatterns in a recursion was greater
1024        than the maximum at the outer level, the higher number was returned, but
1025        with unset values at the outer level. The correct (outer level) value is
1026        now given.
1027    
1028    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1029        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1030        make it do so. The string so far is captured, making this feature
1031        compatible with Perl.
1032    
1033    23. The tests have been re-organized, adding tests 11 and 12, to make it
1034        possible to check the Perl 5.10 features against Perl 5.10.
1035    
1036    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1037        pattern matches a fixed length string. PCRE did not allow this; now it
1038        does. Neither allows recursion.
1039    
1040    25. I finally figured out how to implement a request to provide the minimum
1041        length of subject string that was needed in order to match a given pattern.
1042        (It was back references and recursion that I had previously got hung up
1043        on.) This code has now been added to pcre_study(); it finds a lower bound
1044        to the length of subject needed. It is not necessarily the greatest lower
1045        bound, but using it to avoid searching strings that are too short does give
1046        some useful speed-ups. The value is available to calling programs via
1047        pcre_fullinfo().
1048    
1049    26. While implementing 25, I discovered to my embarrassment that pcretest had
1050        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1051        study optimizations had never been tested with that matching function.
1052        Oops. What is worse, even when it was passed study data, there was a bug in
1053        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1054        were also very few tests of studied patterns with pcre_dfa_exec().
1055    
1056    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1057        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1058        on the other side of the coin, they are no longer allowed to have different
1059        names, because these cannot be distinguished in PCRE, and this has caused
1060        confusion. (This is a difference from Perl.)
1061    
1062    28. When duplicate subpattern names are present (necessarily with different
1063        numbers, as required by 27 above), and a test is made by name in a
1064        conditional pattern, either for a subpattern having been matched, or for
1065        recursion in such a pattern, all the associated numbered subpatterns are
1066        tested, and the overall condition is true if the condition is true for any
1067        one of them. This is the way Perl works, and is also more like the way
1068        testing by number works.
1069    
1070    
1071    Version 7.9 11-Apr-09
1072    ---------------------
1073    
1074    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1075        (pcretest), all targets were linked against these libraries. This included
1076        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1077        libraries. This caused unwanted dependencies to be created. This problem
1078        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1079        pcretest is linked with readline.
1080    
1081    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1082        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1083        moved outside it again, because FALSE and TRUE are already defined in AIX,
1084        but BOOL is not.
1085    
1086    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1087        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1088    
1089    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1090        hyphens) following file names and line numbers when outputting matching
1091        lines. This is not true; no spaces are inserted. I have also clarified the
1092        wording for the --colour (or --color) option.
1093    
1094    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1095        was not coloured; this is different to GNU grep, so I have changed it to be
1096        the same.
1097    
1098    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1099        each matching line was coloured. Now it goes on to look for further matches
1100        of any of the test patterns, which is the same behaviour as GNU grep.
1101    
1102    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1103        doesn't make sense to accept an empty string match in pcregrep, so I have
1104        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1105        seems to be how GNU grep behaves.
1106    
1107    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1108        start or after a newline", because the conditional assertion was not being
1109        correctly handled. The rule now is that both the assertion and what follows
1110        in the first alternative must satisfy the test.
1111    
1112    9.  If auto-callout was enabled in a pattern with a conditional group whose
1113        condition was an assertion, PCRE could crash during matching, both with
1114        pcre_exec() and pcre_dfa_exec().
1115    
1116    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1117        used for matching.
1118    
1119    11. Unicode property support in character classes was not working for
1120        characters (bytes) greater than 127 when not in UTF-8 mode.
1121    
1122    12. Added the -M command line option to pcretest.
1123    
1124    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1125    
1126    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1127    
1128    16. Added comments and documentation about mis-use of no_arg in the C++
1129        wrapper.
1130    
1131    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1132        from Martin Jerabek that uses macro names for all relevant character and
1133        string constants.
1134    
1135    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1136        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1137        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1138        these, but not everybody uses configure.
1139    
1140    19. A conditional group that had only one branch was not being correctly
1141        recognized as an item that could match an empty string. This meant that an
1142        enclosing group might also not be so recognized, causing infinite looping
1143        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1144        with the subject "ab", where knowledge that the repeated group can match
1145        nothing is needed in order to break the loop.
1146    
1147    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1148        exec(), but without supplying a callout function, matching went wrong.
1149    
1150    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1151        leak if the size of the offset vector was greater than 30. When the vector
1152        is smaller, the saved offsets during recursion go onto a local stack
1153        vector, but for larger vectors malloc() is used. It was failing to free
1154        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1155        error, in fact).
1156    
1157    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1158        heapframe that is used only when UTF-8 support is enabled. This caused no
1159        problem, but was untidy.
1160    
1161    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1162        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1163        included within another project.
1164    
1165    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1166        slightly modified by me:
1167    
1168          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1169              not building pcregrep.
1170    
1171          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1172              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1173    
1174    25. Forward references, both numeric and by name, in patterns that made use of
1175        duplicate group numbers, could behave incorrectly or give incorrect errors,
1176        because when scanning forward to find the reference group, PCRE was not
1177        taking into account the duplicate group numbers. A pattern such as
1178        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1179    
1180    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1181        the feature of a custom pointer more persuasive (as requested by a user).
1182    
1183    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1184        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1185        pre-defined.
1186    
1187    28. Added support for (*UTF8) at the start of a pattern.
1188    
1189    29. Arrange for flags added by the "release type" setting in CMake to be shown
1190        in the configuration summary.
1191    
1192    
1193    Version 7.8 05-Sep-08
1194    ---------------------
1195    
1196    1.  Replaced UCP searching code with optimized version as implemented for Ad
1197        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1198        stage table and inline lookup instead of a function, giving speed ups of 2
1199        to 5 times on some simple patterns that I tested. Permission was given to
1200        distribute the MultiStage2.py script that generates the tables (it's not in
1201        the tarball, but is in the Subversion repository).
1202    
1203    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1204        scripts.
1205    
1206    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1207        a group with a zero qualifier. The result of the study could be incorrect,
1208        or the function might crash, depending on the pattern.
1209    
1210    4.  Caseless matching was not working for non-ASCII characters in back
1211        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1212        It now works when Unicode Property Support is available.
1213    
1214    5.  In pcretest, an escape such as \x{de} in the data was always generating
1215        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1216        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1217        truncation.
1218    
1219    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1220    
1221    7.  Added two (int) casts to pcregrep when printing the difference of two
1222        pointers, in case they are 64-bit values.
1223    
1224    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1225        test 2 if it fails.
1226    
1227    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1228        and a #define of that name to empty if it is not externally set. This is to
1229        allow users of MSVC to set it if necessary.
1230    
1231    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1232        the convenience functions in the pcre_get.c source file.
1233    
1234    11. An option change at the start of a pattern that had top-level alternatives
1235        could cause overwriting and/or a crash. This command provoked a crash in
1236        some environments:
1237    
1238          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1239    
1240        This potential security problem was recorded as CVE-2008-2371.
1241    
1242    12. For a pattern where the match had to start at the beginning or immediately
1243        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1244        pcre_dfa_exec() could read past the end of the passed subject if there was
1245        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1246        pcretest so that it places the subject at the end of its malloc-ed buffer.
1247    
1248    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1249        exec() might read past the end of the data buffer in UTF-8 mode.
1250    
1251    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1252        the data contained the byte 0x85 as part of a UTF-8 character within its
1253        first line. This applied both to normal and DFA matching.
1254    
1255    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1256        /^[^d]*?$/8 failed to match "abc".
1257    
1258    16. Added a missing copyright notice to pcrecpp_internal.h.
1259    
1260    17. Make it more clear in the documentation that values returned from
1261        pcre_exec() in ovector are byte offsets, not character counts.
1262    
1263    18. Tidied a few places to stop certain compilers from issuing warnings.
1264    
1265    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1266        supplied by Stefan Weber. I made a further small update for 7.8 because
1267        there is a change of source arrangements: the pcre_searchfuncs.c module is
1268        replaced by pcre_ucd.c.
1269    
1270    
1271    Version 7.7 07-May-08
1272    ---------------------
1273    
1274    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1275        a string to a long long, pretend we don't even have a long long." This is
1276        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1277    
1278    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1279        pre-7.6 versions, which defined a global no_arg variable instead of putting
1280        it in the RE class. (See also #8 below.)
1281    
1282    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1283        Lopes.
1284    
1285    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1286    
1287        (1) The include/exclude patterns were being applied to the whole pathnames
1288            of files, instead of just to the final components.
1289    
1290        (2) If there was more than one level of directory, the subdirectories were
1291            skipped unless they satisfied the include/exclude conditions. This is
1292            inconsistent with GNU grep (and could even be seen as contrary to the
1293            pcregrep specification - which I improved to make it absolutely clear).
1294            The action now is always to scan all levels of directory, and just
1295            apply the include/exclude patterns to regular files.
1296    
1297    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1298        --exclude_dir in the tests to avoid scanning .svn directories.
1299    
1300    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1301        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1302        doesn't support NULs in patterns.
1303    
1304    7.  Added some missing "const"s to declarations of static tables in
1305        pcre_compile.c and pcre_dfa_exec.c.
1306    
1307    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1308        caused by fix #2  above. (Subsequently also a second patch to fix the
1309        first patch. And a third patch - this was a messy problem.)
1310    
1311    9.  Applied Craig's patch to remove the use of push_back().
1312    
1313    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1314        matching function regexec().
1315    
1316    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1317        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1318        references. PCRE supports relative numbers with this syntax (I don't think
1319        Oniguruma does).
1320    
1321    12. Previously, a group with a zero repeat such as (...){0} was completely
1322        omitted from the compiled regex. However, this means that if the group
1323        was called as a subroutine from elsewhere in the pattern, things went wrong
1324        (an internal error was given). Such groups are now left in the compiled
1325        pattern, with a new opcode that causes them to be skipped at execution
1326        time.
1327    
1328    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1329        to the way PCRE behaves:
1330    
1331        (a) A lone ] character is dis-allowed (Perl treats it as data).
1332    
1333        (b) A back reference to an unmatched subpattern matches an empty string
1334            (Perl fails the current match path).
1335    
1336        (c) A data ] in a character class must be notated as \] because if the
1337            first data character in a class is ], it defines an empty class. (In
1338            Perl it is not possible to have an empty class.) The empty class []
1339            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1340            The negative empty class [^] matches any one character, independently
1341            of the DOTALL setting.
1342    
1343    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1344        non-existent subpattern following a character class starting with ']' and
1345        containing () gave an internal compiling error instead of "reference to
1346        non-existent subpattern". Fortunately, when the pattern did exist, the
1347        compiled code was correct. (When scanning forwards to check for the
1348        existencd of the subpattern, it was treating the data ']' as terminating
1349        the class, so got the count wrong. When actually compiling, the reference
1350        was subsequently set up correctly.)
1351    
1352    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1353        it was being rejected as not supported by pcre_dfa_exec(), even though
1354        other assertions are supported. I have made pcre_dfa_exec() support
1355        (*FAIL).
1356    
1357    16. The implementation of 13c above involved the invention of a new opcode,
1358        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1359        cannot be changed at match time, I realized I could make a small
1360        improvement to matching performance by compiling OP_ALLANY instead of
1361        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1362        on the OP_ANY path.
1363    
1364    17. Compiling pcretest on Windows with readline support failed without the
1365        following two fixes: (1) Make the unistd.h include conditional on
1366        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1367    
1368    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1369        ncurses library to be included for pcretest when ReadLine support is
1370        requested, but also to allow for it to be overridden. This patch came from
1371        Daniel Bergström.
1372    
1373    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1374        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1375        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1376        spotting this.
1377    
1378    
1379    Version 7.6 28-Jan-08
1380    ---------------------
1381    
1382    1.  A character class containing a very large number of characters with
1383        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1384        overflow.
1385    
1386    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1387        HAVE_LONG_LONG is not defined.
1388    
1389    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1390        bring them up to date and include new features. This patch includes:
1391    
1392        - Fixed PH's badly added libz and libbz2 support.
1393        - Fixed a problem with static linking.
1394        - Added pcredemo. [But later removed - see 7 below.]
1395        - Fixed dftables problem and added an option.
1396        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1397            HAVE_LONG_LONG.
1398        - Added readline support for pcretest.
1399        - Added an listing of the option settings after cmake has run.
1400    
1401    4.  A user submitted a patch to Makefile that makes it easy to create
1402        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1403        Makefile.am that cause it to include this special target, without
1404        affecting anything else. Note that the same mingw target plus all
1405        the other distribution libraries and programs are now supported
1406        when configuring with CMake (see 6 below) instead of with
1407        Configure/Make.
1408    
1409    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1410        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1411        exported in the Windows port". It has not yet been confirmed that the patch
1412        solves the problem, but it does no harm.
1413    
1414    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1415        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1416        with CMake, and also correct the comment about stack recursion.
1417    
1418    7.  Remove the automatic building of pcredemo from the ./configure system and
1419        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1420        of a program that users should build themselves after PCRE is installed, so
1421        building it automatically is not really right. What is more, it gave
1422        trouble in some build environments.
1423    
1424    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1425    
1426    
1427    Version 7.5 10-Jan-08
1428    ---------------------
1429    
1430    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1431        values in parens when parsing an RE using the C++ wrapper."
1432    
1433    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1434        Characters greater than 255 were excluded from the class instead of being
1435        included.
1436    
1437    3.  The same bug as (2) above applied to negated POSIX classes such as
1438        [:^space:].
1439    
1440    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1441        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1442        I have changed it.
1443    
1444    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1445        first named subpattern) and a construct such as (?&a) would reference the
1446        first named subpattern whose name started with "a" (in other words, the
1447        length check was missing). Both these problems are fixed. "Subpattern name
1448        expected" is now given for (?&) (a zero-length name), and this patch also
1449        makes it give the same error for \k'' (previously it complained that that
1450        was a reference to a non-existent subpattern).
1451    
1452    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1453        this is right because (?- can be followed by option settings as well as by
1454        digits. I have, however, made the messages clearer.
1455    
1456    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1457        than the number used in the conditional) now cause a compile-time error.
1458        This is actually not compatible with Perl, which accepts such patterns, but
1459        treats the conditional as always being FALSE (as PCRE used to), but it
1460        seems to me that giving a diagnostic is better.
1461    
1462    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1463        and messages.
1464    
1465    9.  Fix two occurrences of "backslash" in comments that should have been
1466        "backspace".
1467    
1468    10. Remove two redundant lines of code that can never be obeyed (their function
1469        was moved elsewhere).
1470    
1471    11. The program that makes PCRE's Unicode character property table had a bug
1472        which caused it to generate incorrect table entries for sequences of
1473        characters that have the same character type, but are in different scripts.
1474        It amalgamated them into a single range, with the script of the first of
1475        them. In other words, some characters were in the wrong script. There were
1476        thirteen such cases, affecting characters in the following ranges:
1477    
1478          U+002b0 - U+002c1
1479          U+0060c - U+0060d
1480          U+0061e - U+00612
1481          U+0064b - U+0065e
1482          U+0074d - U+0076d
1483          U+01800 - U+01805
1484          U+01d00 - U+01d77
1485          U+01d9b - U+01dbf
1486          U+0200b - U+0200f
1487          U+030fc - U+030fe
1488          U+03260 - U+0327f
1489          U+0fb46 - U+0fbb1
1490          U+10450 - U+1049d
1491    
1492    12. The -o option (show only the matching part of a line) for pcregrep was not
1493        compatible with GNU grep in that, if there was more than one match in a
1494        line, it showed only the first of them. It now behaves in the same way as
1495        GNU grep.
1496    
1497    13. If the -o and -v options were combined for pcregrep, it printed a blank
1498        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1499        does the same. The return code can be used to tell if there were any
1500        non-matching lines.
1501    
1502    14. Added --file-offsets and --line-offsets to pcregrep.
1503    
1504    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1505        infinitely looping recursion. The bug was that positive lookaheads were not
1506        being skipped when checking for a possible empty match (negative lookaheads
1507        and both kinds of lookbehind were skipped).
1508    
1509    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1510        inclusion of <windows.h> to before rather than after the definition of
1511        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1512    
1513    17. Specifying a possessive quantifier with a specific limit for a Unicode
1514        character property caused pcre_compile() to compile bad code, which led at
1515        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1516        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1517        caused the error; without that there was no problem.
1518    
1519    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1520    
1521    19. Added --enable-pcretest-libreadline.
1522    
1523    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1524        RE::GlobalReplace(). As a result, the number of replacements returned was
1525        double what it should be. I removed one of the increments, but Craig sent a
1526        later patch that removed the other one (the right fix) and added unit tests
1527        that check the return values (which was not done before).
1528    
1529    21. Several CMake things:
1530    
1531        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1532            the names libpcre and libpcreposix, not just pcre and pcreposix.
1533    
1534        (2) The above change means that pcretest and pcregrep are now correctly
1535            linked with the newly-built libraries, not previously installed ones.
1536    
1537        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1538    
1539    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1540        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1541        UTF-8 newline character). The key issue is that the pattern starts .*;
1542        this means that the match must be either at the beginning, or after a
1543        newline. The bug was in the code for advancing after a failed match and
1544        checking that the new position followed a newline. It was not taking
1545        account of UTF-8 characters correctly.
1546    
1547    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1548        character classes. PCRE was not treating the sequence [:...:] as a
1549        character class unless the ... were all letters. Perl, however, seems to
1550        allow any characters between [: and :], though of course it rejects as
1551        unknown any "names" that contain non-letters, because all the known class
1552        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1553        for example, whereas PCRE did not - it did not recognize a POSIX character
1554        class. This seemed a bit dangerous, so the code has been changed to be
1555        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1556        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1557        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1558        Perl does, and where it didn't before.
1559    
1560    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1561        Windows environments %n is disabled by default.
1562    
1563    
1564    Version 7.4 21-Sep-07
1565    ---------------------
1566    
1567    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1568        means that a class such as [\s] counted as "explicit reference to CR or
1569        LF". That isn't really right - the whole point of the change was to try to
1570        help when there was an actual mention of one of the two characters. So now
1571        the change happens only if \r or \n (or a literal CR or LF) character is
1572        encountered.
1573    
1574    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1575        of both had grown to the point where there were only 3 bits left.
1576        Fortunately, there was spare space in the data structure, and so I have
1577        moved the internal flags into a new 16-bit field to free up more option
1578        bits.
1579    
1580    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1581        but did not set the internal JCHANGED flag - either of these is enough to
1582        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1583        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1584        start sets both bits.
1585    
1586    4.  Added options (at build time, compile time, exec time) to change \R from
1587        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1588    
1589    5.  doc/pcresyntax.html was missing from the distribution.
1590    
1591    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1592        compatibility, even though it is no longer used.
1593    
1594    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1595        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1596        windows.h file is present (where different names are used). [This was
1597        reversed later after testing - see 16 below.]
1598    
1599    8.  Changed all #include <config.h> to #include "config.h". There were also
1600        some further <pcre.h> cases that I changed to "pcre.h".
1601    
1602    9.  When pcregrep was used with the --colour option, it missed the line ending
1603        sequence off the lines that it output.
1604    
1605    10. It was pointed out to me that arrays of string pointers cause lots of
1606        relocations when a shared library is dynamically loaded. A technique of
1607        using a single long string with a table of offsets can drastically reduce
1608        these. I have refactored PCRE in four places to do this. The result is
1609        dramatic:
1610    
1611          Originally:                          290
1612          After changing UCP table:            187
1613          After changing error message table:   43
1614          After changing table of "verbs"       36
1615          After changing table of Posix names   22
1616    
1617        Thanks to the folks working on Gregex for glib for this insight.
1618    
1619    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1620        unicode-properties was also set.
1621    
1622    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1623    
1624    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1625        checked only for CRLF.
1626    
1627    14. Added casts to pcretest.c to avoid compiler warnings.
1628    
1629    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1630    
1631    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1632        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1633        entirely. This removes changes made in 7 above.
1634    
1635    17. The CMake files have been updated, and there is now more information about
1636        building with CMake in the NON-UNIX-USE document.
1637    
1638    
1639    Version 7.3 28-Aug-07
1640    ---------------------
1641    
1642     1. In the rejigging of the build system that eventually resulted in 7.1, the
1643        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1644        brackets there is not right, since it causes compilers to look for an
1645        installed pcre.h, not the version that is in the source that is being
1646        compiled (which of course may be different). I have changed it back to:
1647    
1648          #include "pcre.h"
1649    
1650        I have a vague recollection that the change was concerned with compiling in
1651        different directories, but in the new build system, that is taken care of
1652        by the VPATH setting the Makefile.
1653    
1654     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1655        when the subject happened to end in the byte 0x85 (e.g. if the last
1656        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1657        characters but of course it shouldn't be taken as a newline when it is part
1658        of another character. The bug was that, for an unlimited repeat of . in
1659        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1660        characters when looking for a newline.
1661    
1662     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1663    
1664     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1665        in debug output.
1666    
1667     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1668        long printing in the pcrecpp unittest when running under MinGW.
1669    
1670     6. ESC_K was left out of the EBCDIC table.
1671    
1672     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1673        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1674        limit also applies to "virtual nesting" when a pattern is recursive, and in
1675        this case 1000 isn't so big. I have been able to remove this limit at the
1676        expense of backing off one optimization in certain circumstances. Normally,
1677        when pcre_exec() would call its internal match() function recursively and
1678        immediately return the result unconditionally, it uses a "tail recursion"
1679        feature to save stack. However, when a subpattern that can match an empty
1680        string has an unlimited repetition quantifier, it no longer makes this
1681        optimization. That gives it a stack frame in which to save the data for
1682        checking that an empty string has been matched. Previously this was taken
1683        from the 1000-entry workspace that had been reserved. So now there is no
1684        explicit limit, but more stack is used.
1685    
1686     8. Applied Daniel's patches to solve problems with the import/export magic
1687        syntax that is required for Windows, and which was going wrong for the
1688        pcreposix and pcrecpp parts of the library. These were overlooked when this
1689        problem was solved for the main library.
1690    
1691     9. There were some crude static tests to avoid integer overflow when computing
1692        the size of patterns that contain repeated groups with explicit upper
1693        limits. As the maximum quantifier is 65535, the maximum group length was
1694        set at 30,000 so that the product of these two numbers did not overflow a
1695        32-bit integer. However, it turns out that people want to use groups that
1696        are longer than 30,000 bytes (though not repeat them that many times).
1697        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1698        made it possible to implement the integer overflow checks in a much more
1699        dynamic way, which I have now done. The artificial limitation on group
1700        length has been removed - we now have only the limit on the total length of
1701        the compiled pattern, which depends on the LINK_SIZE setting.
1702    
1703    10. Fixed a bug in the documentation for get/copy named substring when
1704        duplicate names are permitted. If none of the named substrings are set, the
1705        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1706        empty string.
1707    
1708    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1709        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1710        because the ] is interpreted as the first data character and the
1711        terminating ] is not found. PCRE has been made compatible with Perl in this
1712        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1713        cause memory overwriting.
1714    
1715    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1716        string has been matched (to stop an infinite loop). It was not recognizing
1717        a conditional subpattern that could match an empty string if that
1718        subpattern was within another subpattern. For example, it looped when
1719        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1720        condition was not nested. This bug has been fixed.
1721    
1722    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1723        past the start of the subject in the presence of bytes with the top bit
1724        set, for example "\x8aBCD".
1725    
1726    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1727        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1728    
1729    14. Optimized (?!) to (*FAIL).
1730    
1731    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1732        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1733        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1734        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1735        does: it's just the validity check that is more restrictive.
1736    
1737    16. Inserted checks for integer overflows during escape sequence (backslash)
1738        processing, and also fixed erroneous offset values for syntax errors during
1739        backslash processing.
1740    
1741    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1742        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1743    
1744    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1745        caused an overrun.
1746    
1747    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1748        something other than just ASCII characters) inside a group that had an
1749        unlimited repeat caused a loop at compile time (while checking to see
1750        whether the group could match an empty string).
1751    
1752    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1753        [\P{Any}] did so. (Error in the code for printing property names.)
1754    
1755    21. An orphan \E inside a character class could cause a crash.
1756    
1757    22. A repeated capturing bracket such as (A)? could cause a wild memory
1758        reference during compilation.
1759    
1760    23. There are several functions in pcre_compile() that scan along a compiled
1761        expression for various reasons (e.g. to see if it's fixed length for look
1762        behind). There were bugs in these functions when a repeated \p or \P was
1763        present in the pattern. These operators have additional parameters compared
1764        with \d, etc, and these were not being taken into account when moving along
1765        the compiled data. Specifically:
1766    
1767        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1768            length.
1769    
1770        (b) An item such as \pL+ within a repeated group could cause crashes or
1771            loops.
1772    
1773        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1774            "reference to non-existent subpattern" error.
1775    
1776        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1777    
1778    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1779        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1780    
1781    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1782    
1783    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1784        character were causing crashes (broken optimization).
1785    
1786    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1787        \p or \P) caused a compile-time loop.
1788    
1789    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1790        break. For example, the unstudied pattern [\r\n]A does not match the string
1791        "\r\nA" because change 7.0/46 below moves the current point on by two
1792        characters after failing to match at the start. However, the pattern \nA
1793        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1794        the same is true. There doesn't seem any very clean way out of this, but
1795        what I have chosen to do makes the common cases work: PCRE now takes note
1796        of whether there can be an explicit match for \r or \n anywhere in the
1797        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1798        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1799        pattern has explicit CR or LF references.
1800    
1801    29. Added (*CR) etc for changing newline setting at start of pattern.
1802    
1803    
1804    Version 7.2 19-Jun-07
1805    ---------------------
1806    
1807     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1808        which is apparently normally available under Windows.
1809    
1810     2. Re-jig the pcregrep tests with different newline settings in an attempt
1811        to make them independent of the local environment's newline setting.
1812    
1813     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1814    
1815     4. Some of the "internals" tests were previously cut out when the link size
1816        was not 2, because the output contained actual offsets. The recent new
1817        "Z" feature of pcretest means that these can be cut out, making the tests
1818        usable with all link sizes.
1819    
1820     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1821        stack recursion. This gives a massive performance boost under BSD, but just
1822        a small improvement under Linux. However, it saves one field in the frame
1823        in all cases.
1824    
1825     6. Added more features from the forthcoming Perl 5.10:
1826    
1827        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1828            recursion call. It refers to the nth most recently opened parentheses.
1829    
1830        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1831            to be opened parentheses.
1832    
1833        (c) Conditions that refer to capturing parentheses can be specified
1834            relatively, for example, (?(-2)... or (?(+3)...
1835    
1836        (d) \K resets the start of the current match so that everything before
1837            is not part of it.
1838    
1839        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1840    
1841        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1842            reference syntax.
1843    
1844        (g) (?| introduces a group in which the numbering of parentheses in each
1845            alternative starts with the same number.
1846    
1847        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1848    
1849     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1850        PCRE_INFO_JCHANGED.
1851    
1852     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1853        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1854        for detecting groups that can match an empty string.
1855    
1856     9. A pattern with a very large number of alternatives (more than several
1857        hundred) was running out of internal workspace during the pre-compile
1858        phase, where pcre_compile() figures out how much memory will be needed. A
1859        bit of new cunning has reduced the workspace needed for groups with
1860        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1861        workspace instead of running out of the 4096 that are available.
1862    
1863    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1864    
1865    11. Applied patch from Google to remove an optimization that didn't quite work.
1866        The report of the bug said:
1867    
1868          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1869          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1870          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1871    
1872    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1873        it matched the wrong number of bytes.
1874    
1875    
1876    Version 7.1 24-Apr-07
1877    ---------------------
1878    
1879     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1880        that is more "standard", making use of automake and other Autotools. There
1881        is some re-arrangement of the files and adjustment of comments consequent
1882        on this.
1883    
1884     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1885        for recursive directory scanning broke on some systems because the files
1886        are not scanned in any specific order and on different systems the order
1887        was different. A call to "sort" has been inserted into RunGrepTest for the
1888        approprate test as a short-term fix. In the longer term there may be an
1889        alternative.
1890    
1891     3. I had an email from Eric Raymond about problems translating some of PCRE's
1892        man pages to HTML (despite the fact that I distribute HTML pages, some
1893        people do their own conversions for various reasons). The problems
1894        concerned the use of low-level troff macros .br and .in. I have therefore
1895        removed all such uses from the man pages (some were redundant, some could
1896        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1897        HTML has been updated to handle .nf/.fi and to complain if it encounters
1898        .br or .in.
1899    
1900     4. Updated comments in configure.ac that get placed in config.h.in and also
1901        arranged for config.h to be included in the distribution, with the name
1902        config.h.generic, for the benefit of those who have to compile without
1903        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1904    
1905     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1906        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1907        makevp.bat for the current PCRE, using the additional files
1908        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1909    
1910     6. A Windows user reported a minor discrepancy with test 2, which turned out
1911        to be caused by a trailing space on an input line that had got lost in his
1912        copy. The trailing space was an accident, so I've just removed it.
1913    
1914     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1915        that is needed.
1916    
1917     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1918        as "const" (a) because they are and (b) because it helps the PHP
1919        maintainers who have recently made a script to detect big data structures
1920        in the php code that should be moved to the .rodata section. I remembered
1921        to update Builducptable as well, so it won't revert if ucptable.h is ever
1922        re-created.
1923    
1924     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1925        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1926        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1927        support is not required. This saves 1.5-2K of code, which is important in
1928        some applications.
1929    
1930        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1931        so as not to refer to the tables, even though these functions will never be
1932        called when UTF-8 support is disabled. Otherwise there are problems with a
1933        shared library.
1934    
1935    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1936    
1937        (a) It was defining its arguments as char * instead of void *.
1938    
1939        (b) It was assuming that all moves were upwards in memory; this was true
1940            a long time ago when I wrote it, but is no longer the case.
1941    
1942        The emulated memove() is provided for those environments that have neither
1943        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1944        is clearly not the case, as these two bugs were recently reported.
1945    
1946    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1947        and Detrail to create the HTML documentation, the .txt form of the man
1948        pages, and it removes trailing spaces from listed files. It also creates
1949        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1950        case, it wraps all the #defines with #ifndefs. This script should be run
1951        before "make dist".
1952    
1953    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1954        with Unicode property support.
1955    
1956        (a) For a maximizing quantifier, if the two different cases of the
1957            character were of different lengths in their UTF-8 codings (there are
1958            some cases like this - I found 11), and the matching function had to
1959            back up over a mixture of the two cases, it incorrectly assumed they
1960            were both the same length.
1961    
1962        (b) When PCRE was configured to use the heap rather than the stack for
1963            recursion during matching, it was not correctly preserving the data for
1964            the other case of a UTF-8 character when checking ahead for a match
1965            while processing a minimizing repeat. If the check also involved
1966            matching a wide character, but failed, corruption could cause an
1967            erroneous result when trying to check for a repeat of the original
1968            character.
1969    
1970    13. Some tidying changes to the testing mechanism:
1971    
1972        (a) The RunTest script now detects the internal link size and whether there
1973            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1974            values substituted by "configure". (The RunGrepTest script already did
1975            this for UTF-8.) The configure.ac script no longer substitutes the
1976            relevant variables.
1977    
1978        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1979            with length and offset values. This means that the output is different
1980            for different internal link sizes. Test 2 is skipped for link sizes
1981            other than 2 because of this, bypassing the problem. Unfortunately,
1982            there was also a test in test 3 (the locale tests) that used /B and
1983            failed for link sizes other than 2. Rather than cut the whole test out,
1984            I have added a new /Z option to pcretest that replaces the length and
1985            offset values with spaces. This is now used to make test 3 independent
1986            of link size. (Test 2 will be tidied up later.)
1987    
1988    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1989        segmentation fault instead of returning the appropriate error message.
1990    
1991    15. In multiline mode when the newline sequence was set to "any", the pattern
1992        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1993        This doesn't seem right; it now treats the CRLF combination as the line
1994        ending, and so does not match in that case. It's only a pattern such as ^$
1995        that would hit this one: something like ^ABC$ would have failed after \r
1996        and then tried again after \r\n.
1997    
1998    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1999        in an attempt to make files that differ only in their line terminators
2000        compare equal. This works on Linux.
2001    
2002    17. Under certain error circumstances pcregrep might try to free random memory
2003        as it exited. This is now fixed, thanks to valgrind.
2004    
2005    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2006        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2007        was because its rules for how to advance for /g after matching an empty
2008        string at the end of a line did not allow for this case. They now check for
2009        it specially.
2010    
2011    20. pcretest is supposed to handle patterns and data of any length, by
2012        extending its buffers when necessary. It was getting this wrong when the
2013        buffer for a data line had to be extended.
2014    
2015    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2016        CRLF as a newline sequence.
2017    
2018    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2019        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2020        I have nevertheless tidied it up.
2021    
2022    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2023    
2024    24. Added a man page for pcre-config.
2025    
2026    
2027    Version 7.0 19-Dec-06
2028    ---------------------
2029    
2030     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2031        moving to gcc 4.1.1.
2032    
2033     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2034        sys/time.h, which is documented as needed for this function. It doesn't
2035        seem to matter on Linux, but it showed up on some releases of OS X.
2036    
2037     3. It seems that there are systems where bytes whose values are greater than
2038        127 match isprint() in the "C" locale. The "C" locale should be the
2039        default when a C program starts up. In most systems, only ASCII printing
2040        characters match isprint(). This difference caused the output from pcretest
2041        to vary, making some of the tests fail. I have changed pcretest so that:
2042    
2043        (a) When it is outputting text in the compiled version of a pattern, bytes
2044            other than 32-126 are always shown as hex escapes.
2045    
2046        (b) When it is outputting text that is a matched part of a subject string,
2047            it does the same, unless a different locale has been set for the match
2048            (using the /L modifier). In this case, it uses isprint() to decide.
2049    
2050     4. Fixed a major bug that caused incorrect computation of the amount of memory
2051        required for a compiled pattern when options that changed within the
2052        pattern affected the logic of the preliminary scan that determines the
2053        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2054        that the computed length was too small. The symptoms of this bug were
2055        either the PCRE error "internal error: code overflow" from pcre_compile(),
2056        or a glibc crash with a message such as "pcretest: free(): invalid next
2057        size (fast)". Examples of patterns that provoked this bug (shown in
2058        pcretest format) are:
2059    
2060          /(?-x: )/x
2061          /(?x)(?-x: \s*#\s*)/
2062          /((?i)[\x{c0}])/8
2063          /(?i:[\x{c0}])/8
2064    
2065        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2066        is now done differently.
2067    
2068     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2069        wrapper classes; (b) implement a new function in the C++ scanner that is
2070        more efficient than the old way of doing things because it avoids levels of
2071        recursion in the regex matching; (c) add a paragraph to the documentation
2072        for the FullMatch() function.
2073    
2074     6. The escape sequence \n was being treated as whatever was defined as
2075        "newline". Not only was this contrary to the documentation, which states
2076        that \n is character 10 (hex 0A), but it also went horribly wrong when
2077        "newline" was defined as CRLF. This has been fixed.
2078    
2079     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2080        was being set to -1 for the "end of line" case (supposedly a value that no
2081        character can have). Though this value is never used (the check for end of
2082        line is "zero bytes in current character"), it caused compiler complaints.
2083        I've changed it to 0xffffffff.
2084    
2085     8. In pcre_version.c, the version string was being built by a sequence of
2086        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2087        string (as it is for production releases) called a macro with an empty
2088        argument. The C standard says the result of this is undefined. The gcc
2089        compiler treats it as an empty string (which was what was wanted) but it is
2090        reported that Visual C gives an error. The source has been hacked around to
2091        avoid this problem.
2092    
2093     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2094        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2095        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2096        of them did).
2097    
2098    10. Originally, pcretest opened its input and output without "b"; then I was
2099        told that "b" was needed in some environments, so it was added for release
2100        5.0 to both the input and output. (It makes no difference on Unix-like
2101        systems.) Later I was told that it is wrong for the input on Windows. I've
2102        now abstracted the modes into two macros, to make it easier to fiddle with
2103        them, and removed "b" from the input mode under Windows.
2104    
2105    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2106    
2107    12. Added -help and --help to pcretest as an official way of being reminded
2108        of the options.
2109    
2110    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2111        and pcrecpp.cc because they annoy compilers at high warning levels.
2112    
2113    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2114    
2115    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2116        scripts are not C programs :-) and which was not noticed because it works
2117        on Linux.
2118    
2119    16. pcretest is supposed to handle any length of pattern and data line (as one
2120        line or as a continued sequence of lines) by extending its input buffer if
2121        necessary. This feature was broken for very long pattern lines, leading to
2122        a string of junk being passed to pcre_compile() if the pattern was longer
2123        than about 50K.
2124    
2125    17. I have done a major re-factoring of the way pcre_compile() computes the
2126        amount of memory needed for a compiled pattern. Previously, there was code
2127        that made a preliminary scan of the pattern in order to do this. That was
2128        OK when PCRE was new, but as the facilities have expanded, it has become
2129        harder and harder to keep it in step with the real compile phase, and there
2130        have been a number of bugs (see for example, 4 above). I have now found a
2131        cunning way of running the real compile function in a "fake" mode that
2132        enables it to compute how much memory it would need, while actually only
2133        ever using a few hundred bytes of working memory and without too many
2134        tests of the mode. This should make future maintenance and development
2135        easier. A side effect of this work is that the limit of 200 on the nesting
2136        depth of parentheses has been removed (though this was never a serious
2137        limitation, I suspect). However, there is a downside: pcre_compile() now
2138        runs more slowly than before (30% or more, depending on the pattern). I
2139        hope this isn't a big issue. There is no effect on runtime performance.
2140    
2141    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2142        newline (only possible for the last line of a file) and it was a
2143        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2144    
2145    19. Added additional timing features to pcretest. (1) The -tm option now times
2146        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2147        separate command line item, by a number that specifies the number of
2148        repeats to use when timing. The default is 50000; this gives better
2149        precision, but takes uncomfortably long for very large patterns.
2150    
2151    20. Extended pcre_study() to be more clever in cases where a branch of a
2152        subpattern has no definite first character. For example, (a*|b*)[cd] would
2153        previously give no result from pcre_study(). Now it recognizes that the
2154        first character must be a, b, c, or d.
2155    
2156    21. There was an incorrect error "recursive call could loop indefinitely" if
2157        a subpattern (or the entire pattern) that was being tested for matching an
2158        empty string contained only one non-empty item after a nested subpattern.
2159        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2160        incorrectly, because the \d was being skipped in the check.
2161    
2162    22. The pcretest program now has a new pattern option /B and a command line
2163        option -b, which is equivalent to adding /B to every pattern. This causes
2164        it to show the compiled bytecode, without the additional information that
2165        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2166        is the same as /B/I).
2167    
2168    23. A new optimization is now able automatically to treat some sequences such
2169        as a*b as a*+b. More specifically, if something simple (such as a character
2170        or a simple class like \d) has an unlimited quantifier, and is followed by
2171        something that cannot possibly match the quantified thing, the quantifier
2172        is automatically "possessified".
2173    
2174    24. A recursive reference to a subpattern whose number was greater than 39
2175        went wrong under certain circumstances in UTF-8 mode. This bug could also
2176        have affected the operation of pcre_study().
2177    
2178    25. Realized that a little bit of performance could be had by replacing
2179        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2180    
2181    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2182    
2183    27. Possessive quantifiers such as a++ were previously implemented by turning
2184        them into atomic groups such as ($>a+). Now they have their own opcodes,
2185        which improves performance. This includes the automatically created ones
2186        from 23 above.
2187    
2188    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2189        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2190        the first matched character to be a colon. This applied both to named and
2191        numbered groups.
2192    
2193    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2194    
2195    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2196        building PCRE on Windows easier, so I have included it in the distribution.
2197    
2198    31. There is now a check in pcretest against a ridiculously large number being
2199        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2200        loop, the loop is abandoned.
2201    
2202    32. Forward references to subpatterns in conditions such as (?(2)...) where
2203        subpattern 2 is defined later cause pcre_compile() to search forwards in
2204        the pattern for the relevant set of parentheses. This search went wrong
2205        when there were unescaped parentheses in a character class, parentheses
2206        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2207    
2208    33. "Subroutine" calls and backreferences were previously restricted to
2209        referencing subpatterns earlier in the regex. This restriction has now
2210        been removed.
2211    
2212    34. Added a number of extra features that are going to be in Perl 5.10. On the
2213        whole, these are just syntactic alternatives for features that PCRE had
2214        previously implemented using the Python syntax or my own invention. The
2215        other formats are all retained for compatibility.
2216    
2217        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2218            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2219            also .NET compatible.
2220    
2221        (b) A recursion or subroutine call to a named group can now be defined as
2222            (?&name) as well as (?P>name).
2223    
2224        (c) A backreference to a named group can now be defined as \k<name> or
2225            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2226            5.10, are also .NET compatible.
2227    
2228        (d) A conditional reference to a named group can now use the syntax
2229            (?(<name>) or (?('name') as well as (?(name).
2230    
2231        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2232            groups (named and numbered) that are never evaluated inline, but can be
2233            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2234            is always false. There may be only one alternative in such a group.
2235    
2236        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2237            as the simple (?(R). The condition is true only if the most recent
2238            recursion is that of the given number or name. It does not search out
2239            through the entire recursion stack.
2240    
2241        (g) The escape \gN or \g{N} has been added, where N is a positive or
2242            negative number, specifying an absolute or relative reference.
2243    
2244    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2245        some "unreachable code" warnings.
2246    
2247    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2248        things, this adds five new scripts.
2249    
2250    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2251        There were also incompatibilities regarding the handling of \Q..\E inside
2252        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2253        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2254    
2255    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2256        matches an empty string, and forcibly breaks the loop. There were bugs in
2257        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2258        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2259        separate and independent bugs (that affected different cases) have been
2260        fixed.
2261    
2262    39. Refactored the code to abolish the use of different opcodes for small
2263        capturing bracket numbers. This is a tidy that I avoided doing when I
2264        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2265        The new approach is not only tidier, it makes it possible to reduce the
2266        memory needed to fix the previous bug (38).
2267    
2268    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2269        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2270        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2271        mode.
2272    
2273    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2274        report.
2275    
2276    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2277        copy construction and assignment in the C++ wrapper.
2278    
2279    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2280        couple of bugs that could have given wrong results in the "--newline=crlf"
2281        case.
2282    
2283    44. Added a number of casts and did some reorganization of signed/unsigned int
2284        variables following suggestions from Dair Grant. Also renamed the variable
2285        "this" as "item" because it is a C++ keyword.
2286    
2287    45. Arranged for dftables to add
2288    
2289          #include "pcre_internal.h"
2290    
2291        to pcre_chartables.c because without it, gcc 4.x may remove the array
2292        definition from the final binary if PCRE is built into a static library and
2293        dead code stripping is activated.
2294    
2295    46. For an unanchored pattern, if a match attempt fails at the start of a
2296        newline sequence, and the newline setting is CRLF or ANY, and the next two
2297        characters are CRLF, advance by two characters instead of one.
2298    
2299    
2300    Version 6.7 04-Jul-06
2301    ---------------------
2302    
2303     1. In order to handle tests when input lines are enormously long, pcretest has
2304        been re-factored so that it automatically extends its buffers when
2305        necessary. The code is crude, but this _is_ just a test program. The
2306        default size has been increased from 32K to 50K.
2307    
2308     2. The code in pcre_study() was using the value of the re argument before
2309        testing it for NULL. (Of course, in any sensible call of the function, it
2310        won't be NULL.)
2311    
2312     3. The memmove() emulation function in pcre_internal.h, which is used on
2313        systems that lack both memmove() and bcopy() - that is, hardly ever -
2314        was missing a "static" storage class specifier.
2315    
2316     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2317        containing an extended class (one that cannot be represented by a bitmap
2318        because it contains high-valued characters or Unicode property items, e.g.
2319        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2320        pattern, but PCRE should not loop if you do not (it no longer does).
2321        [Detail: two cases were found: (a) a repeated subpattern containing an
2322        extended class; (b) a recursive reference to a subpattern that followed a
2323        previous extended class. It wasn't skipping over the extended class
2324        correctly when UTF-8 mode was not set.]
2325    
2326     5. A negated single-character class was not being recognized as fixed-length
2327        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2328        compile error "lookbehind assertion is not fixed length".
2329    
2330     6. The RunPerlTest auxiliary script was showing an unexpected difference
2331        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2332        write a Perl script that can interpret lines of an input file either as
2333        byte characters or as UTF-8, which is what "perltest" was being required to
2334        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2335        can't do is switch easily at run time between having the "use utf8;" pragma
2336        or not. In the end, I fudged it by using the RunPerlTest script to insert
2337        "use utf8;" explicitly for the UTF-8 tests.
2338    
2339     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2340        the end of the subject string, contrary to the documentation and to what
2341        Perl does. This was true of both matching functions. Now it matches only at
2342        the start of the subject and immediately after *internal* newlines.
2343    
2344     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2345        a pointer to an int instead of a pointer to an unsigned long int. This
2346        caused problems on 64-bit systems.
2347    
2348     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2349        instance of the 'standard' template library not being so standard".
2350    
2351    10. There was no check on the number of named subpatterns nor the maximum
2352        length of a subpattern name. The product of these values is used to compute
2353        the size of the memory block for a compiled pattern. By supplying a very
2354        long subpattern name and a large number of named subpatterns, the size
2355        computation could be caused to overflow. This is now prevented by limiting
2356        the length of names to 32 characters, and the number of named subpatterns
2357        to 10,000.
2358    
2359    11. Subpatterns that are repeated with specific counts have to be replicated in
2360        the compiled pattern. The size of memory for this was computed from the
2361        length of the subpattern and the repeat count. The latter is limited to
2362        65535, but there was no limit on the former, meaning that integer overflow
2363        could in principle occur. The compiled length of a repeated subpattern is
2364        now limited to 30,000 bytes in order to prevent this.
2365    
2366    12. Added the optional facility to have named substrings with the same name.
2367    
2368    13. Added the ability to use a named substring as a condition, using the
2369        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2370        are numbers (not recommended). Forward references are permitted.
2371    
2372    14. Added forward references in named backreferences (if you see what I mean).
2373    
2374    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2375        pattern could run off the end of the subject. For example, the pattern
2376        "(?s)(.{1,5})"8 did this with the subject "ab".
2377    
2378    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2379        PCRE_CASELESS was set when matching characters that were quantified with ?
2380        or *.
2381    
2382    17. A character class other than a single negated character that had a minimum
2383        but no maximum quantifier - for example [ab]{6,} - was not handled
2384        correctly by pce_dfa_exec(). It would match only one character.
2385    
2386    18. A valid (though odd) pattern that looked like a POSIX character
2387        class but used an invalid character after [ (for example [[,abc,]]) caused
2388        pcre_compile() to give the error "Failed: internal error: code overflow" or
2389        in some cases to crash with a glibc free() error. This could even happen if
2390        the pattern terminated after [[ but there just happened to be a sequence of
2391        letters, a binary zero, and a closing ] in the memory that followed.
2392    
2393    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2394        over the years. Originally (before any Unicode support), just the bottom 8
2395        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2396        output from "man perlunicode" includes this:
2397    
2398          The regular expression compiler produces polymorphic opcodes.  That
2399          is, the pattern adapts to the data and automatically switches to
2400          the Unicode character scheme when presented with Unicode data--or
2401          instead uses a traditional byte scheme when presented with byte
2402          data.
2403    
2404        Sadly, a wide octal escape does not cause a switch, and in a string with
2405        no other multibyte characters, these octal escapes are treated as before.
2406        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2407        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2408        Unicode string.
2409    
2410        I have not perpetrated such confusion in PCRE. Up till now, it took just
2411        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2412        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2413        translate to the appropriate multibyte character.
2414    
2415    29. Applied some refactoring to reduce the number of warnings from Microsoft
2416        and Borland compilers. This has included removing the fudge introduced
2417        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2418        a warning about an unused variable.
2419    
2420    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2421        characters since release 4.0, because Perl (from release 5.004) does not.
2422        [Or at least, is documented not to: some releases seem to be in conflict
2423        with the documentation.] However, when a pattern was studied with
2424        pcre_study() and all its branches started with \s, PCRE still included VT
2425        as a possible starting character. Of course, this did no harm; it just
2426        caused an unnecessary match attempt.
2427    
2428    22. Removed a now-redundant internal flag bit that recorded the fact that case
2429        dependency changed within the pattern. This was once needed for "required
2430        byte" processing, but is no longer used. This recovers a now-scarce options
2431        bit. Also moved the least significant internal flag bit to the most-
2432        significant bit of the word, which was not previously used (hangover from
2433        the days when it was an int rather than a uint) to free up another bit for
2434        the future.
2435    
2436    23. Added support for CRLF line endings as well as CR and LF. As well as the
2437        default being selectable at build time, it can now be changed at runtime
2438        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2439        specify that it is scanning data with non-default line endings.
2440    
2441    24. Changed the definition of CXXLINK to make it agree with the definition of
2442        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2443    
2444    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2445        recursions. This makes a big different to stack usage for some patterns.
2446    
2447    26. If a subpattern containing a named recursion or subroutine reference such
2448        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2449        the space required for the compiled pattern went wrong and gave too small a
2450        value. Depending on the environment, this could lead to "Failed: internal
2451        error: code overflow at offset 49" or "glibc detected double free or
2452        corruption" errors.
2453    
2454    27. Applied patches from Google (a) to support the new newline modes and (b) to
2455        advance over multibyte UTF-8 characters in GlobalReplace.
2456    
2457    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2458        difference for some implementation of PCRE in some Windows version.
2459    
2460    29. Added some extra testing facilities to pcretest:
2461    
2462        \q<number>   in a data line sets the "match limit" value
2463        \Q<number>   in a data line sets the "match recursion limt" value
2464        -S <number>  sets the stack size, where <number> is in megabytes
2465    
2466        The -S option isn't available for Windows.
2467    
2468    
2469    Version 6.6 06-Feb-06
2470    ---------------------
2471    
2472     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2473        in pcreposix.h. I have copied the definition from pcre.h.
2474    
2475     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2476        because pcre.h is no longer a built file.
2477    
2478     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2479        not normally included in the compiled code.
2480    
2481    
2482    Version 6.5 01-Feb-06
2483    ---------------------
2484    
2485     1. When using the partial match feature with pcre_dfa_exec(), it was not
2486        anchoring the second and subsequent partial matches at the new starting
2487        point. This could lead to incorrect results. For example, with the pattern
2488        /1234/, partially matching against "123" and then "a4" gave a match.
2489    
2490     2. Changes to pcregrep:
2491    
2492        (a) All non-match returns from pcre_exec() were being treated as failures
2493            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2494            error message is output. Some extra information is given for the
2495            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2496            probably the only errors that are likely to be caused by users (by
2497            specifying a regex that has nested indefinite repeats, for instance).
2498            If there are more than 20 of these errors, pcregrep is abandoned.
2499    
2500        (b) A binary zero was treated as data while matching, but terminated the
2501            output line if it was written out. This has been fixed: binary zeroes
2502            are now no different to any other data bytes.
2503    
2504        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2505            used to set a locale for matching. The --locale=xxxx long option has
2506            been added (no short equivalent) to specify a locale explicitly on the
2507            pcregrep command, overriding the environment variables.
2508    
2509        (d) When -B was used with -n, some line numbers in the output were one less
2510            than they should have been.
2511    
2512        (e) Added the -o (--only-matching) option.
2513    
2514        (f) If -A or -C was used with -c (count only), some lines of context were
2515            accidentally printed for the final match.
2516    
2517        (g) Added the -H (--with-filename) option.
2518    
2519        (h) The combination of options -rh failed to suppress file names for files
2520            that were found from directory arguments.
2521    
2522        (i) Added the -D (--devices) and -d (--directories) options.
2523    
2524        (j) Added the -F (--fixed-strings) option.
2525    
2526        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2527    
2528        (l) Added the --colo(u)r option.
2529    
2530        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2531            is not present by default.
2532    
2533     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2534        items such as (?R) or (?1), when the recursion could match a number of
2535        alternatives. If it matched one of the alternatives, but subsequently,
2536        outside the recursion, there was a failure, the code tried to back up into
2537        the recursion. However, because of the way PCRE is implemented, this is not
2538        possible, and the result was an incorrect result from the match.
2539    
2540        In order to prevent this happening, the specification of recursion has
2541        been changed so that all such subpatterns are automatically treated as
2542        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2543    
2544     4. I had overlooked the fact that, in some locales, there are characters for
2545        which isalpha() is true but neither isupper() nor islower() are true. In
2546        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2547        and ordfeminine) are like this. This affected the treatment of \w and \W
2548        when they appeared in character classes, but not when they appeared outside
2549        a character class. The bit map for "word" characters is now created
2550        separately from the results of isalnum() instead of just taking it from the
2551        upper, lower, and digit maps. (Plus the underscore character, of course.)
2552    
2553     5. The above bug also affected the handling of POSIX character classes such as
2554        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2555        permanent tables. Instead, the bit maps for such a class were previously
2556        created as the appropriate unions of the upper, lower, and digit bitmaps.
2557        Now they are created by subtraction from the [[:word:]] class, which has
2558        its own bitmap.
2559    
2560     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2561        It is created by subtracting the vertical space characters (\x09, \x0a,
2562        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2563        subtraction was done in the overall bitmap for a character class, meaning
2564        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2565        be recognized. This bug has been fixed.
2566    
2567     7. Patches from the folks at Google:
2568    
2569          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2570          real life, but is still worth protecting against".
2571    
2572          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2573          regular expressions".
2574    
2575          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2576          have it.
2577    
2578          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2579          "configure" and the latter not, in order to fix a problem somebody had
2580          with compiling the Arg class on HP-UX.
2581    
2582          (e) Improve the error-handling of the C++ wrapper a little bit.
2583    
2584          (f) New tests for checking recursion limiting.
2585    
2586     8. The pcre_memmove() function, which is used only if the environment does not
2587        have a standard memmove() function (and is therefore rarely compiled),
2588        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2589        returning a result (though PCRE never actually uses the result).
2590    
2591     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2592        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2593        returned instead of calling malloc() with an overflowing number that would
2594        most likely cause subsequent chaos.
2595    
2596    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2597    
2598    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2599        with this option is matched, the nmatch and pmatch options of regexec() are
2600        ignored.
2601    
2602    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2603        provided in case anyone wants to the the POSIX interface with UTF-8
2604        strings.
2605    
2606    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2607        C++ linking (needed for some HP-UX environments).
2608    
2609    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2610        (unused parameter) and in the pcre_printint() function (omitted "default"
2611        switch label when the default is to do nothing).
2612    
2613    15. Added some code to make it possible, when PCRE is compiled as a C++
2614        library, to replace subject pointers for pcre_exec() with a smart pointer
2615        class, thus making it possible to process discontinuous strings.
2616    
2617    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2618        much the same function. They were added by different people who were trying
2619        to make PCRE easy to compile on non-Unix systems. It has been suggested
2620        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2621        for compiling on Windows systems. I have therefore replaced it with
2622        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2623        defaults to "extern" for C or "extern C" for C++, which works fine on
2624        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2625        SCOPE with something explicit in config.h. In addition:
2626    
2627        (a) pcreposix.h still had just "extern" instead of either of these macros;
2628            I have replaced it with PCRE_DATA_SCOPE.
2629    
2630        (b) Functions such as _pcre_xclass(), which are internal to the library,
2631            but external in the C sense, all had PCRE_EXPORT in their definitions.
2632            This is apparently wrong for the Windows case, so I have removed it.
2633            (It makes no difference on Unix-like systems.)
2634    
2635    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2636        of recursive calls to match(). This is different to MATCH_LIMIT because
2637        that limits the total number of calls to match(), not all of which increase
2638        the depth of recursion. Limiting the recursion depth limits the amount of
2639        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2640        when PCRE is compiled, and changed at run time. A patch from Google adds
2641        this functionality to the C++ interface.
2642    
2643    18. Changes to the handling of Unicode character properties:
2644    
2645        (a) Updated the table to Unicode 4.1.0.
2646    
2647        (b) Recognize characters that are not in the table as "Cn" (undefined).
2648    
2649        (c) I revised the way the table is implemented to a much improved format
2650            which includes recognition of ranges. It now supports the ranges that
2651            are defined in UnicodeData.txt, and it also amalgamates other
2652            characters into ranges. This has reduced the number of entries in the
2653            table from around 16,000 to around 3,000, thus reducing its size
2654            considerably. I realized I did not need to use a tree structure after
2655            all - a binary chop search is just as efficient. Having reduced the
2656            number of entries, I extended their size from 6 bytes to 8 bytes to
2657            allow for more data.
2658    
2659        (d) Added support for Unicode script names via properties such as \p{Han}.
2660    
2661    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2662        matching that character.
2663    
2664    20. When matching a repeated Unicode property with a minimum greater than zero,
2665        (for example \pL{2,}), PCRE could look past the end of the subject if it
2666        reached it while seeking the minimum number of characters. This could
2667        happen only if some of the characters were more than one byte long, because
2668        there is a check for at least the minimum number of bytes.
2669    
2670    21. Refactored the implementation of \p and \P so as to be more general, to
2671        allow for more different types of property in future. This has changed the
2672        compiled form incompatibly. Anybody with saved compiled patterns that use
2673        \p or \P will have to recompile them.
2674    
2675    22. Added "Any" and "L&" to the supported property types.
2676    
2677    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2678        but give a compile time error if the value is greater than 0xff.
2679    
2680    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2681        accidentally not being installed or uninstalled.
2682    
2683    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2684        made were to insert the current release number. This seemed silly, because
2685        it made things harder for people building PCRE on systems that don't run
2686        "configure". I have turned pcre.h into a distributed file, no longer built
2687        by "configure", with the version identification directly included. There is
2688        no longer a pcre.h.in file.
2689    
2690        However, this change necessitated a change to the pcre-config script as
2691        well. It is built from pcre-config.in, and one of the substitutions was the
2692        release number. I have updated configure.ac so that ./configure now finds
2693        the release number by grepping pcre.h.
2694    
2695    26. Added the ability to run the tests under valgrind.
2696    
2697    
2698    Version 6.4 05-Sep-05
2699    ---------------------
2700    
2701     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2702        "--" to be printed when multiple files were scanned, even when none of the
2703        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2704        consider it to be a bug, and have restored the previous behaviour.
2705    
2706     2. A couple of code tidies to get rid of compiler warnings.
2707    
2708     3. The pcretest program used to cheat by referring to symbols in the library
2709        whose names begin with _pcre_. These are internal symbols that are not
2710        really supposed to be visible externally, and in some environments it is
2711        possible to suppress them. The cheating is now confined to including
2712        certain files from the library's source, which is a bit cleaner.
2713    
2714     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2715        file's purpose clearer.
2716    
2717     5. Reorganized pcre_ucp_findchar().
2718    
2719    
2720    Version 6.3 15-Aug-05
2721    ---------------------
2722    
2723     1. The file libpcre.pc.in did not have general read permission in the tarball.
2724    
2725     2. There were some problems when building without C++ support:
2726    
2727        (a) If C++ support was not built, "make install" and "make test" still
2728            tried to test it.
2729    
2730        (b) There were problems when the value of CXX was explicitly set. Some
2731            changes have been made to try to fix these, and ...
2732    
2733        (c) --disable-cpp can now be used to explicitly disable C++ support.
2734    
2735        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2736            backslash in a target when C++ was disabled. This confuses some
2737            versions of "make", apparently. Using an intermediate variable solves
2738            this. (Same for CPP_LOBJ.)
2739    
2740     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2741        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2742        necessary on certain architectures.
2743    
2744     4. Added a setting of -export-symbols-regex to the link command to remove
2745        those symbols that are exported in the C sense, but actually are local
2746        within the library, and not documented. Their names all begin with
2747        "_pcre_". This is not a perfect job, because (a) we have to except some
2748        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2749        available (and never for static libraries). I have made a note to try to
2750        find a way round (a) in the future.
2751    
2752    
2753    Version 6.2 01-Aug-05
2754    ---------------------
2755    
2756     1. There was no test for integer overflow of quantifier values. A construction
2757        such as {1111111111111111} would give undefined results. What is worse, if
2758        a minimum quantifier for a parenthesized subpattern overflowed and became
2759        negative, the calculation of the memory size went wrong. This could have
2760        led to memory overwriting.
2761    
2762     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2763    
2764     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2765        operating environments where this matters.
2766    
2767     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2768        PCRE options from within the C++ wrapper.
2769    
2770     5. Named capturing subpatterns were not being correctly counted when a pattern
2771        was compiled. This caused two problems: (a) If there were more than 100
2772        such subpatterns, the calculation of the memory needed for the whole
2773        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2774        back references of the form \12, where the number was greater than 9, were
2775        not recognized as back references, even though there were sufficient
2776        previous subpatterns.
2777    
2778     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2779        versions of gcc, e.g. 2.95.4.
2780    
2781    
2782    Version 6.1 21-Jun-05
2783    ---------------------
2784    
2785     1. There was one reference to the variable "posix" in pcretest.c that was not
2786        surrounded by "#if !defined NOPOSIX".
2787    
2788     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2789        the cross-check on the old pcre_info() function, for the benefit of the
2790        cut-down version of PCRE that is currently imported into Exim.
2791    
2792     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2793        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2794        patterns that start (?i) but I don't think that matters. The use of (?i) is
2795        just an example; this all applies to the other options as well.
2796    
2797     4. Since libtool seems to echo the compile commands it is issuing, the output
2798        from "make" can be reduced a bit by putting "@" in front of each libtool
2799        compile command.
2800    
2801     5. Patch from the folks at Google for configure.in to be a bit more thorough
2802        in checking for a suitable C++ installation before trying to compile the
2803        C++ stuff. This should fix a reported problem when a compiler was present,
2804        but no suitable headers.
2805    
2806     6. The man pages all had just "PCRE" as their title. I have changed them to
2807        be the relevant file name. I have also arranged that these names are
2808        retained in the file doc/pcre.txt, which is a concatenation in text format
2809        of all the man pages except the little individual ones for each function.
2810    
2811     7. The NON-UNIX-USE file had not been updated for the different set of source
2812        files that come with release 6. I also added a few comments about the C++
2813        wrapper.
2814    
2815    
2816    Version 6.0 07-Jun-05
2817    ---------------------
2818    
2819     1. Some minor internal re-organization to help with my DFA experiments.
2820    
2821     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2822        didn't matter for the library itself when fully configured, but did matter
2823        when compiling without UCP support, or within Exim, where the ucp files are
2824        not imported.
2825    
2826     3. Refactoring of the library code to split up the various functions into
2827        different source modules. The addition of the new DFA matching code (see
2828        below) to a single monolithic source would have made it really too
2829        unwieldy, quite apart from causing all the code to be include in a
2830        statically linked application, when only some functions are used. This is
2831        relevant even without the DFA addition now that patterns can be compiled in
2832        one application and matched in another.
2833    
2834        The downside of splitting up is that there have to be some external
2835        functions and data tables that are used internally in different modules of
2836        the library but which are not part of the API. These have all had their
2837        names changed to start with "_pcre_" so that they are unlikely to clash
2838        with other external names.
2839    
2840     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2841        a different (DFA) algorithm. Although it is slower than the original
2842        function, it does have some advantages for certain types of matching
2843        problem.
2844    
2845     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2846        including restarting after a partial match.
2847    
2848     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2849        defined when compiling for Windows was sent to me. I have put it into the
2850        code, though I have no means of testing or verifying it.
2851    
2852     7. Added the pcre_refcount() auxiliary function.
2853    
2854     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2855        match before or at the first newline in the subject string. In pcretest,
2856        the /f option on a pattern can be used to set this.
2857    
2858     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2859        would behave wrongly. This has been present in PCRE since release 4.0.
2860    
2861    10. A number of changes to the pcregrep command:
2862    
2863        (a) Refactored how -x works; insert ^(...)$ instead of setting
2864            PCRE_ANCHORED and checking the length, in preparation for adding
2865            something similar for -w.
2866    
2867        (b) Added the -w (match as a word) option.
2868    
2869        (c) Refactored the way lines are read and buffered so as to have more
2870            than one at a time available.
2871    
2872        (d) Implemented a pcregrep test script.
2873    
2874        (e) Added the -M (multiline match) option. This allows patterns to match
2875            over several lines of the subject. The buffering ensures that at least
2876            8K, or the rest of the document (whichever is the shorter) is available
2877            for matching (and similarly the previous 8K for lookbehind assertions).
2878    
2879        (f) Changed the --help output so that it now says
2880    
2881              -w, --word-regex(p)
2882    
2883            instead of two lines, one with "regex" and the other with "regexp"
2884            because that confused at least one person since the short forms are the
2885            same. (This required a bit of code, as the output is generated
2886            automatically from a table. It wasn't just a text change.)
2887    
2888        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2889            option but starts with a hyphen. Could be a pattern or a path name
2890            starting with a hyphen, for instance.
2891    
2892        (h) "-" can be given as a file name to represent stdin.
2893    
2894        (i) When file names are being printed, "(standard input)" is used for
2895            the standard input, for compatibility with GNU grep. Previously
2896            "<stdin>" was used.
2897    
2898        (j) The option --label=xxx can be used to supply a name to be used for
2899            stdin when file names are being printed. There is no short form.
2900    
2901        (k) Re-factored the options decoding logic because we are going to add
2902            two more options that take data. Such options can now be given in four
2903            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2904    
2905        (l) Added the -A, -B, and -C options for requesting that lines of context
2906            around matches be printed.
2907    
2908        (m) Added the -L option to print the names of files that do not contain
2909            any matching lines, that is, the complement of -l.
2910    
2911        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2912            continue to scan other files.
2913    
2914        (o) The -s option was incorrectly implemented. For compatibility with other
2915            greps, it now suppresses the error message for a non-existent or non-
2916            accessible file (but not the return code). There is a new option called
2917            -q that suppresses the output of matching lines, which was what -s was
2918            previously doing.
2919    
2920        (p) Added --include and --exclude options to specify files for inclusion
2921            and exclusion when recursing.
2922    
2923    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2924        Hopefully, it now does.
2925    
2926    12. Missing cast in pcre_study().
2927    
2928    13. Added an "uninstall" target to the makefile.
2929    
2930    14. Replaced "extern" in the function prototypes in Makefile.in with
2931        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2932        world, but is set differently for Windows.
2933    
2934    15. Added a second compiling function called pcre_compile2(). The only
2935        difference is that it has an extra argument, which is a pointer to an
2936        integer error code. When there is a compile-time failure, this is set
2937        non-zero, in addition to the error test pointer being set to point to an
2938        error message. The new argument may be NULL if no error number is required
2939        (but then you may as well call pcre_compile(), which is now just a
2940        wrapper). This facility is provided because some applications need a
2941        numeric error indication, but it has also enabled me to tidy up the way
2942        compile-time errors are handled in the POSIX wrapper.
2943    
2944    16. Added VPATH=.libs to the makefile; this should help when building with one
2945        prefix path and installing with another. (Or so I'm told by someone who
2946        knows more about this stuff than I do.)
2947    
2948    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2949        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2950        match everything, including newlines. This is not POSIX-compatible, but
2951        somebody wanted the feature. From pcretest it can be activated by using
2952        both the P and the s flags.
2953    
2954    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2955    
2956    19. libpcre.pc was being incorrectly installed as executable.
2957    
2958    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2959        it now also looks for '\r' so that it will work unmodified on Windows.
2960    
2961    21. Added Google's contributed C++ wrapper to the distribution.
2962    
2963    22. Added some untidy missing memory free() calls in pcretest, to keep
2964        Electric Fence happy when testing.
2965    
2966    
2967    
2968    Version 5.0 13-Sep-04
2969    ---------------------
2970    
2971     1. Internal change: literal characters are no longer packed up into items
2972        containing multiple characters in a single byte-string. Each character
2973        is now matched using a separate opcode. However, there may be more than one
2974        byte in the character in UTF-8 mode.
2975    
2976     2. The pcre_callout_block structure has two new fields: pattern_position and
2977        next_item_length. These contain the offset in the pattern to the next match
2978        item, and its length, respectively.
2979    
2980     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2981        insertion of callouts before each pattern item. Added the /C option to
2982        pcretest to make use of this.
2983    
2984     4. On the advice of a Windows user, the lines
2985    
2986          #if defined(_WIN32) || defined(WIN32)
2987          _setmode( _fileno( stdout ), 0x8000 );
2988          #endif  /* defined(_WIN32) || defined(WIN32) */
2989    
2990        have been added to the source of pcretest. This apparently does useful
2991        magic in relation to line terminators.
2992    
2993     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2994        for the benefit of those environments where the "b" makes a difference.
2995    
2996     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2997        to know about it. I have put a hack into configure.in that adds in code
2998        to set GCC=yes if CC=icc. This seems to end up at a point in the
2999        generated configure script that is early enough to affect the setting of
3000        compiler options, which is what is needed, but I have no means of testing
3001        whether it really works. (The user who reported this had patched the
3002        generated configure script, which of course I cannot do.)
3003    
3004        LATER: After change 22 below (new libtool files), the configure script
3005        seems to know about icc (and also ecc). Therefore, I have commented out
3006        this hack in configure.in.
3007    
3008     7. Added support for pkg-config (2 patches were sent in).
3009    
3010     8. Negated POSIX character classes that used a combination of internal tables
3011        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3012        [[:^ascii]]. Typically, they would match almost any characters. The other
3013        POSIX classes were not broken in this way.
3014    
3015     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3016        to find the match, as PCRE was deluded into thinking that the match had to
3017        start at the start point or following a newline. The same bug applied to
3018        patterns with negative forward assertions or any backward assertions
3019        preceding ".*" at the start, unless the pattern required a fixed first
3020        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3021    
3022    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3023        starting at the last subject character, bytes beyond the end of the subject
3024        string were read.
3025    
3026    11. Renamed the variable "class" as "classbits" to make life easier for C++
3027        users. (Previously there was a macro definition, but it apparently wasn't
3028        enough.)
3029    
3030    12. Added the new field "tables" to the extra data so that tables can be passed
3031        in at exec time, or the internal tables can be re-selected. This allows
3032        a compiled regex to be saved and re-used at a later time by a different
3033        program that might have everything at different addresses.
3034    
3035    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3036        -R library as well as a -L library.
3037    
3038    14. The debugging options of pcretest (-d on the command line or D on a
3039        pattern) showed incorrect output for anything following an extended class
3040        that contained multibyte characters and which was followed by a quantifier.
3041    
3042    15. Added optional support for general category Unicode character properties
3043        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3044        support. It adds about 90K to the size of the library. The meanings of the
3045        inbuilt class escapes such as \d and \s have NOT been changed.
3046    
3047    16. Updated pcredemo.c to include calls to free() to release the memory for the
3048        compiled pattern.
3049    
3050    17. The generated file chartables.c was being created in the source directory
3051        instead of in the building directory. This caused the build to fail if the
3052        source directory was different from the building directory, and was
3053        read-only.
3054    
3055    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3056        file. No doubt somebody will tell me if they don't make sense... Also added
3057        Dan Mooney's comments about building on OpenVMS.
3058    
3059    19. Added support for partial matching via the PCRE_PARTIAL option for
3060        pcre_exec() and the \P data escape in pcretest.
3061    
3062    20. Extended pcretest with 3 new pattern features:
3063    
3064        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3065              write the compiled pattern to the file whose name is "rest-of-line".
3066              This is a straight binary dump of the data, with the saved pointer to
3067              the character tables forced to be NULL. The study data, if any, is
3068              written too. After writing, pcretest reads a new pattern.
3069    
3070        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3071              compiled pattern from the given file. There must not be any
3072              occurrences of "<" in the file name (pretty unlikely); if there are,
3073              pcretest will instead treat the initial "<" as a pattern delimiter.
3074              After reading in the pattern, pcretest goes on to read data lines as
3075              usual.
3076    
3077        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3078              and 16-bit fields in a compiled pattern, to simulate a pattern that
3079              was compiled on a host of opposite endianness.
3080    
3081    21. The pcre-exec() function can now cope with patterns that were compiled on
3082        hosts of opposite endianness, with this restriction:
3083    
3084          As for any compiled expression that is saved and used later, the tables
3085          pointer field cannot be preserved; the extra_data field in the arguments
3086          to pcre_exec() should be used to pass in a tables address if a value
3087          other than the default internal tables were used at compile time.
3088    
3089    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3090        now diagnosed as an error. Previously, most of the time, a negative number
3091        would have been treated as zero, but if in addition "ovector" was passed as
3092        NULL, a crash could occur.
3093    
3094    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3095        new versions from the libtool 1.5 distribution (the last one is a copy of
3096        a file called libtool.m4). This seems to have fixed the need to patch
3097        "configure" to support Darwin 1.3 (which I used to do). However, I still
3098        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3099        workstation).
3100    
3101    24. Changed the PCRE licence to be the more standard "BSD" licence.
3102    
3103    
3104    Version 4.5 01-Dec-03
3105    ---------------------
3106    
3107     1. There has been some re-arrangement of the code for the match() function so
3108        that it can be compiled in a version that does not call itself recursively.
3109        Instead, it keeps those local variables that need separate instances for
3110        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3111        needs to "recurse". Keeping track of where control must go is done by means
3112        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3113        hide most of the details from the main code, and operates only if
3114        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3115        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3116        operating.
3117    
3118        To make it easier for callers to provide specially tailored get/free
3119        functions for this usage, two new functions, pcre_stack_malloc, and
3120        pcre_stack_free, are used. They are always called in strict stacking order,
3121        and the size of block requested is always the same.
3122    
3123        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3124        PCRE has been compiled to use the stack or the heap for recursion. The
3125        -C option of pcretest uses this to show which version is compiled.
3126    
3127        A new data escape \S, is added to pcretest; it causes the amounts of store
3128        obtained and freed by both kinds of malloc/free at match time to be added
3129        to the output.
3130    
3131     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3132        what's available on my current Linux desktop machine.
3133    
3134     3. When matching a UTF-8 string, the test for a valid string at the start has
3135        been extended. If start_offset is not zero, PCRE now checks that it points
3136        to a byte that is the start of a UTF-8 character. If not, it returns
3137        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3138        this is necessary because there may be backward assertions in the pattern.
3139        When matching the same subject several times, it may save resources to use
3140        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3141    
3142     4. The code for checking the validity of UTF-8 strings has been tightened so
3143        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3144        containing "overlong sequences".
3145    
3146     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3147        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3148        should have been "&", but it just so happened that all the cases this let
3149        through by mistake were picked up later in the function.
3150    
3151     6. I had used a variable called "isblank" - this is a C99 function, causing
3152        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3153    
3154     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3155        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3156        so the version is shown; (c) stop "make test" echoing "./RunTest".
3157    
3158     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3159    
3160     9. The prototype for memmove() for systems that don't have it was using
3161        size_t, but the inclusion of the header that defines size_t was later. I've
3162        moved the #includes for the C headers earlier to avoid this.
3163    
3164    10. Added some adjustments to the code to make it easier to compiler on certain
3165        special systems:
3166    
3167          (a) Some "const" qualifiers were missing.
3168          (b) Added the macro EXPORT before all exported functions; by default this
3169              is defined to be empty.
3170          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3171              that it reads its output file name as an argument instead of writing
3172              to the standard output and assuming this can be redirected.
3173    
3174    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3175        class containing characters with values greater than 255, PCRE compilation
3176        went into a loop.
3177    
3178    12. A recursive reference to a subpattern that was within another subpattern
3179        that had a minimum quantifier of zero caused PCRE to crash. For example,
3180        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3181        recursion. If the recursively-called subpattern itself had a zero repeat,
3182        that was OK.
3183    
3184    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3185        buffer into which it was copied (for escape processing) was still set at
3186        1024, so long lines caused crashes.
3187    
3188    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3189        "internal error: code overflow...". This applied to any character class
3190        that was followed by a possessive quantifier.
3191    
3192    15. Modified the Makefile to add libpcre.la as a prerequisite for
3193        libpcreposix.la because I was told this is needed for a parallel build to
3194        work.
3195    
3196    16. If a pattern that contained .* following optional items at the start was
3197        studied, the wrong optimizing data was generated, leading to matching
3198        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3199        matching string must start with a or b or c. The correct conclusion for
3200        this pattern is that a match can start with any character.
3201    
3202    
3203    Version 4.4 13-Aug-03
3204    ---------------------
3205    
3206     1. In UTF-8 mode, a character class containing characters with values between
3207        127 and 255 was not handled correctly if the compiled pattern was studied.
3208        In fixing this, I have also improved the studying algorithm for such
3209        classes (slightly).
3210    
3211     2. Three internal functions had redundant arguments passed to them. Removal
3212        might give a very teeny performance improvement.
3213    
3214     3. Documentation bug: the value of the capture_top field in a callout is *one
3215        more than* the number of the hightest numbered captured substring.
3216    
3217     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3218        in incorrectly linking with a previously installed version. They now link
3219        explicitly with libpcre.la.
3220    
3221     5. configure.in no longer needs to recognize Cygwin specially.
3222    
3223     6. A problem in pcre.in for Windows platforms is fixed.
3224    
3225     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3226        pcretest, it used to include the size of the study block as part of its
3227        output. Unfortunately, the structure contains a field that has a different
3228        size on different hardware architectures. This meant that the tests that
3229        showed this size failed. As the block is currently always of a fixed size,
3230        this information isn't actually particularly useful in pcretest output, so
3231        I have just removed it.
3232    
3233     8. Three pre-processor statements accidentally did not start in column 1.
3234        Sadly, there are *still* compilers around that complain, even though
3235        standard C has not required this for well over a decade. Sigh.
3236    
3237     9. In pcretest, the code for checking callouts passed small integers in the
3238        callout_data field, which is a void * field. However, some picky compilers
3239        complained about the casts involved for this on 64-bit systems. Now
3240        pcretest passes the address of the small integer instead, which should get
3241        rid of the warnings.
3242    
3243    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3244        both compile and run time, and gives an error if an invalid UTF-8 sequence
3245        is found. There is a option for disabling this check in cases where the
3246        string is known to be correct and/or the maximum performance is wanted.
3247    
3248    11. In response to a bug report, I changed one line in Makefile.in from
3249    
3250            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3251        to
3252            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3253    
3254        to look similar to other lines, but I have no way of telling whether this
3255        is the right thing to do, as I do not use Windows. No doubt I'll get told
3256        if it's wrong...
3257    
3258    
3259    Version 4.3 21-May-03
3260    ---------------------
3261    
3262    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3263       Makefile.
3264    
3265    2. Some refactoring to improve the quality of the code:
3266    
3267       (i)   The utf8_table... variables are now declared "const".
3268    
3269       (ii)  The code for \cx, which used the "case flipping" table to upper case
3270             lower case letters, now just substracts 32. This is ASCII-specific,
3271             but the whole concept of \cx is ASCII-specific, so it seems
3272             reasonable.
3273    
3274       (iii) PCRE was using its character types table to recognize decimal and
3275             hexadecimal digits in the pattern. This is silly, because it handles
3276             only 0-9, a-f, and A-F, but the character types table is locale-
3277             specific, which means strange things might happen. A private
3278             table is now used for this - though it costs 256 bytes, a table is
3279             much faster than multiple explicit tests. Of course, the standard
3280             character types table is still used for matching digits in subject
3281             strings against \d.
3282    
3283       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3284             ending in _t are). So I've renamed it as ESC_tee.
3285    
3286    3. The first argument for regexec() in the POSIX wrapper should have been
3287       defined as "const".
3288    
3289    4. Changed pcretest to use malloc() for its buffers so that they can be
3290       Electric Fenced for debugging.
3291    
3292    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3293       to read one or more bytes before the start of the subject string. Often this
3294       had no effect on PCRE's behaviour, but in some circumstances it could
3295       provoke a segmentation fault.
3296    
3297    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3298       to try to read one or more bytes before the start of the subject string.
3299    
3300    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3301       UTF-8 support could misbehave in various ways if the subject string
3302       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3303       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3304       back over UTF-8 characters.)
3305    
3306    
3307    Version 4.2 14-Apr-03
3308    ---------------------
3309    
3310    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3311    
3312    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3313         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3314         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3315         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3316         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3317           and BUILD_EXEEXT
3318         Note: automatic setting of the BUILD variables is not yet working
3319         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3320           compile-time but not at link-time
3321         [LINK]: use for linking executables only
3322         make different versions for Windows and non-Windows
3323         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3324           libraries
3325         [LINK_FOR_BUILD]: new variable
3326         [OBJEXT]: use throughout
3327         [EXEEXT]: use throughout
3328         <winshared>: new target
3329         <wininstall>: new target
3330         <dftables.o>: use native compiler
3331         <dftables>: use native linker
3332         <install>: handle Windows platform correctly
3333         <clean>: ditto
3334         <check>: ditto
3335         copy DLL to top builddir before testing
3336    
3337       As part of these changes, -no-undefined was removed again. This was reported
3338       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3339       in any case.
3340    
3341    3. Some tidies to get rid of compiler warnings:
3342    
3343       . In the match_data structure, match_limit was an unsigned long int, whereas
3344         match_call_count was an int. I've made them both unsigned long ints.
3345    
3346       . In pcretest the fact that a const uschar * doesn't automatically cast to
3347         a void * provoked a warning.
3348    
3349       . Turning on some more compiler warnings threw up some "shadow" variables
3350         and a few more missing casts.
3351    
3352    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3353       option, a class that contained a single character with a value between 128
3354       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3355    
3356    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3357       option, a class that contained several characters, but with at least one
3358       whose value was between 128 and 255 caused PCRE to crash.
3359    
3360    
3361    Version 4.1 12-Mar-03
3362    ---------------------
3363    
3364    1. Compiling with gcc -pedantic found a couple of places where casts were
3365    needed, and a string in dftables.c that was longer than standard compilers are
3366    required to support.
3367    
3368    2. Compiling with Sun's compiler found a few more places where the code could
3369    be tidied up in order to avoid warnings.
3370    
3371    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3372    first of these names is deprecated in the latest Autoconf in favour of the name
3373    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3374    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3375    analogy I have changed it to CFLAGS_FOR_BUILD.
3376    
3377    4. Added -no-undefined to the linking command in the Makefile, because this is
3378    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3379    linking step for the pcreposix library.
3380    
3381    5. PCRE was failing to diagnose the case of two named groups with the same
3382    name.
3383    
3384    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3385    literal character that is needed in the subject for a match, and scans along to
3386    ensure that it is present before embarking on the full matching process. This
3387    saves time in cases of nested unlimited repeats that are never going to match.
3388    Problem: the scan can take a lot of time if the subject is very long (e.g.
3389    megabytes), thus penalizing straightforward matches. It is now done only if the
3390    amount of subject to be scanned is less than 1000 bytes.
3391    
3392    7. A lesser problem with the same optimization is that it was recording the
3393    first character of an anchored pattern as "needed", thus provoking a search
3394    right along the subject, even when the first match of the pattern was going to
3395    fail. The "needed" character is now not set for anchored patterns, unless it
3396    follows something in the pattern that is of non-fixed length. Thus, it still
3397    fulfils its original purpose of finding quick non-matches in cases of nested
3398    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3399    
3400    
3401    Version 4.0 17-Feb-03
3402    ---------------------
3403    
3404    1. If a comment in an extended regex that started immediately after a meta-item
3405    extended to the end of string, PCRE compiled incorrect data. This could lead to
3406    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3407    
3408    2. Moved to autoconf 2.53 and libtool 1.4.2.
3409    
3410    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3411    the special perltest8 script is no longer needed - all the tests can be run
3412    from a single perltest script.
3413    
3414    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3415    by \s. It has now been removed in PCRE. This means it isn't recognized as
3416    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3417    class [:space:] *does* include VT, thereby creating a mess.
3418    
3419    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3420    space and tab.
3421    
3422    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3423    its new features into the main test script, reducing the number of scripts.
3424    
3425    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3426    were backward compatible, and made the (?i) apply to the whole pattern, as if
3427    /i were given. Now it behaves more logically, and applies the option setting
3428    only to what follows. PCRE has been changed to follow suit. However, if it
3429    finds options settings right at the start of the pattern, it extracts them into
3430    the global options, as before. Thus, they show up in the info data.
3431    
3432    8. Added support for the \Q...\E escape sequence. Characters in between are
3433    treated as literals. This is slightly different from Perl in that $ and @ are
3434    also handled as literals inside the quotes. In Perl, they will cause variable
3435    interpolation. Note the following examples:
3436    
3437        Pattern            PCRE matches      Perl matches
3438    
3439        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3440        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3441        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3442    
3443    For compatibility with Perl, \Q...\E sequences are recognized inside character
3444    classes as well as outside them.
3445    
3446    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3447    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3448    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3449    signed/unsigned warnings.
3450    
3451    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3452    option for pcretest, so I've replaced it by a simple function that does just
3453    that job.
3454    
3455    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3456    "pcregrep -".
3457    
3458    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3459    Java package. This provides some syntactic sugar for simple cases of what my
3460    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3461    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3462    item, you can use this simplified notation. Note that only makes sense with
3463    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3464    greediness, whatever the setting of the PCRE_UNGREEDY option.
3465    
3466    13. A change of greediness default within a pattern was not taking effect at
3467    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3468    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3469    was abstracted outside.
3470    
3471    14. PCRE now supports the \G assertion. It is true when the current matching
3472    position is at the start point of the match. This differs from \A when the
3473    starting offset is non-zero. Used with the /g option of pcretest (or similar
3474    code), it works in the same way as it does for Perl's /g option. If all
3475    alternatives of a regex begin with \G, the expression is anchored to the start
3476    match position, and the "anchored" flag is set in the compiled expression.
3477    
3478    15. Some bugs concerning the handling of certain option changes within patterns
3479    have been fixed. These applied to options other than (?ims). For example,
3480    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3481    been the other way round. Some of this was related to change 7 above.
3482    
3483    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3484    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3485    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3486    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3487    
3488    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3489    mode. Unlike ".", it always matches newline, whatever the setting of
3490    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3491    assertions. Perl allows it, but it doesn't (in general) work because it can't
3492    calculate the length of the lookbehind. At least, that's the case for Perl
3493    5.8.0 - I've been told they are going to document that it doesn't work in
3494    future.
3495    
3496    18. Added an error diagnosis for escapes that PCRE does not support: these are
3497    \L, \l, \N, \P, \p, \U, \u, and \X.
3498    
3499    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3500    reading past the end of the pattern in cases such as /[abcd/.
3501    
3502    20. PCRE was getting more memory than necessary for patterns with classes that
3503    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3504    
3505    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3506    compiling PCRE for use with Virtual Pascal.
3507    
3508    22. Small fix to the Makefile to make it work properly if the build is done
3509    outside the source tree.
3510    
3511    23. Added a new extension: a condition to go with recursion. If a conditional
3512    subpattern starts with (?(R) the "true" branch is used if recursion has
3513    happened, whereas the "false" branch is used only at the top level.
3514    
3515    24. When there was a very long string of literal characters (over 255 bytes
3516    without UTF support, over 250 bytes with UTF support), the computation of how
3517    much memory was required could be incorrect, leading to segfaults or other
3518    strange effects.
3519    
3520    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3521    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3522    there was a subsequent back reference to those brackets. This meant that, for
3523    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3524    possible to check for precisely this case. All we can do is abandon the
3525    optimization if .* occurs inside capturing brackets when there are any back
3526    references whatsoever. (See below for a better fix that came later.)
3527    
3528    26. The handling of the optimization for finding the first character of a
3529    non-anchored pattern, and for finding a character that is required later in the
3530    match were failing in some cases. This didn't break the matching; it just
3531    failed to optimize when it could. The way this is done has been re-implemented.
3532    
3533    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3534    
3535    28. Added a new feature that provides some of the functionality that Perl
3536    provides with (?{...}). The facility is termed a "callout". The way it is done
3537    in PCRE is for the caller to provide an optional function, by setting
3538    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3539    global variable. By default it is unset, which disables all calling out. To get
3540    the function called, the regex must include (?C) at appropriate points. This
3541    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3542    This provides a means of identifying different callout points. When PCRE
3543    reaches such a point in the regex, if pcre_callout has been set, the external
3544    function is called. It is provided with data in a structure called
3545    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3546    matching continues; if it returns a non-zero value, the match at the current
3547    point fails. However, backtracking will occur if possible. [This was changed
3548    later and other features added - see item 49 below.]
3549    
3550    29. pcretest is upgraded to test the callout functionality. It provides a
3551    callout function that displays information. By default, it shows the start of
3552    the match and the current position in the text. There are some new data escapes
3553    to vary what happens:
3554    
3555        \C+         in addition, show current contents of captured substrings
3556        \C-         do not supply a callout function
3557        \C!n        return 1 when callout number n is reached
3558        \C!n!m      return 1 when callout number n is reached for the mth time
3559    
3560    30. If pcregrep was called with the -l option and just a single file name, it
3561    output "<stdin>" if a match was found, instead of the file name.
3562    
3563    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3564    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3565    pcre_exec(). This saves a malloc/free per call. The default value of
3566    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3567    when configuring.
3568    
3569    32. The default maximum size of a compiled pattern is 64K. There have been a
3570    few cases of people hitting this limit. The code now uses macros to handle the
3571    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3572    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3573    configuring. Tests 2 and 5 work only with 2-byte links because they output
3574    debugging information about compiled patterns.
3575    
3576    33. Internal code re-arrangements:
3577    
3578    (a) Moved the debugging function for printing out a compiled regex into
3579        its own source file (printint.c) and used #include to pull it into
3580        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3581        separate copies.
3582    
3583    (b) Defined the list of op-code names for debugging as a macro in
3584        internal.h so that it is next to the definition of the opcodes.
3585    
3586    (c) Defined a table of op-code lengths for simpler skipping along compiled
3587        code. This is again a macro in internal.h so that it is next to the
3588        definition of the opcodes.
3589    
3590    34. Added support for recursive calls to individual subpatterns, along the
3591    lines of Robin Houston's patch (but implemented somewhat differently).
3592    
3593    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3594    allow it to read and process whole directories in Win32. This code was
3595    contributed by Lionel Fourquaux; it has not been tested by me.
3596    
3597    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3598    used to name a group. Names consist of alphanumerics and underscores, and must
3599    be unique. Back references use the syntax (?P=name) and recursive calls use
3600    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3601    numbers. The function pcre_fullinfo() can be used after compilation to extract
3602    a name/number map. There are three relevant calls:
3603    
3604      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3605      PCRE_INFO_NAMECOUNT            yields the number of entries
3606      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3607    
3608    The map is a vector of fixed-size entries. The size of each entry depends on
3609    the length of the longest name used. The first two bytes of each entry are the
3610    group number, most significant byte first. There follows the corresponding
3611    name, zero terminated. The names are in alphabetical order.
3612    
3613    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3614    case instead of 255. Making it the same both with and without UTF-8 support
3615    means that the same test output works with both.
3616    
3617    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3618    calling malloc() with a zero argument.
3619    
3620    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3621    optimization. I've improved things by keeping a bitmap of backreferences with
3622    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3623    fact referenced, the optimization can be applied. It is unlikely that a
3624    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3625    the match to follow \n) will appear inside brackets with a number greater than
3626    31, but if it does, any back reference > 31 suppresses the optimization.
3627    
3628    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3629    of disabling numbered capturing parentheses. Any opening parenthesis that is
3630    not followed by ? behaves as if it were followed by ?: but named parentheses
3631    can still be used for capturing (and they will acquire numbers in the usual
3632    way).
3633    
3634    41. Redesigned the return codes from the match() function into yes/no/error so
3635    that errors can be passed back from deep inside the nested calls. A malloc
3636    failure while inside a recursive subpattern call now causes the
3637    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3638    
3639    42. It is now possible to set a limit on the number of times the match()
3640    function is called in a call to pcre_exec(). This facility makes it possible to
3641    limit the amount of recursion and backtracking, though not in a directly
3642    obvious way, because the match() function is used in a number of different
3643    circumstances. The count starts from zero for each position in the subject
3644    string (for non-anchored patterns). The default limit is, for compatibility, a
3645    large number, namely 10 000 000. You can change this in two ways:
3646    
3647    (a) When configuring PCRE before making, you can use --with-match-limit=n
3648        to set a default value for the compiled library.
3649    
3650    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3651        a different value is set. See 45 below.
3652    
3653    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3654    
3655    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3656    of things that can be changed at compile time. The first argument specifies
3657    what is wanted and the second points to where the information is to be placed.
3658    The current list of available information is:
3659    
3660      PCRE_CONFIG_UTF8
3661    
3662    The output is an integer that is set to one if UTF-8 support is available;
3663    otherwise it is set to zero.
3664    
3665      PCRE_CONFIG_NEWLINE
3666    
3667    The output is an integer that it set to the value of the code that is used for
3668    newline. It is either LF (10) or CR (13).
3669    
3670      PCRE_CONFIG_LINK_SIZE
3671    
3672    The output is an integer that contains the number of bytes used for internal
3673    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3674    
3675      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3676    
3677    The output is an integer that contains the threshold above which the POSIX
3678    interface uses malloc() for output vectors. See item 31 above.
3679    
3680      PCRE_CONFIG_MATCH_LIMIT
3681    
3682    The output is an unsigned integer that contains the default limit of the number
3683    of match() calls in a pcre_exec() execution. See 42 above.
3684    
3685    44. pcretest has been upgraded by the addition of the -C option. This causes it
3686    to extract all the available output from the new pcre_config() function, and to
3687    output it. The program then exits immediately.
3688    
3689    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3690    order to support additional features. One way would have been to define
3691    pcre_exec2() (for example) with extra arguments, but this would not have been
3692    extensible, and would also have required all calls to the original function to
3693    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3694    is used for passing in "extra" data from pcre_study().
3695    
3696    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3697    contains the following fields:
3698    
3699      flags         a bitmap indicating which of the following fields are set
3700      study_data    opaque data from pcre_study()
3701      match_limit   a way of specifying a limit on match() calls for a specific
3702                      call to pcre_exec()
3703      callout_data  data for callouts (see 49 below)
3704    
3705    The flag bits are also defined in pcre.h, and are
3706    
3707      PCRE_EXTRA_STUDY_DATA
3708      PCRE_EXTRA_MATCH_LIMIT
3709      PCRE_EXTRA_CALLOUT_DATA
3710    
3711    The pcre_study() function now returns one of these new pcre_extra blocks, with
3712    the actual study data pointed to by the study_data field, and the
3713    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3714    before. That is, this change is entirely upwards-compatible and requires no
3715    change to existing code.
3716    
3717    If you want to pass in additional data to pcre_exec(), you can either place it
3718    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3719    block.
3720    
3721    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3722    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3723    times with different match limits, until it finds the minimum value needed for
3724    pcre_exec() to complete. The value is then output. This can be instructive; for
3725    most simple matches the number is quite small, but for pathological cases it
3726    gets very large very quickly.
3727    
3728    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3729    returns the size of the data block pointed to by the study_data field in a
3730    pcre_extra block, that is, the value that was passed as the argument to
3731    pcre_malloc() when PCRE was getting memory in which to place the information
3732    created by pcre_study(). The fourth argument should point to a size_t variable.
3733    pcretest has been extended so that this information is shown after a successful
3734    pcre_study() call when information about the compiled regex is being displayed.
3735    
3736    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3737    because what follows is always an absolute path. (Later: it turns out that this
3738    is more than cosmetic for MinGW, because it doesn't like empty path
3739    components.)
3740    
3741    49. Some changes have been made to the callout feature (see 28 above):
3742    
3743    (i)  A callout function now has three choices for what it returns:
3744    
3745           0  =>  success, carry on matching
3746         > 0  =>  failure at this point, but backtrack if possible
3747         < 0  =>  serious error, return this value from pcre_exec()
3748    
3749         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3750         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3751         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3752         use by callout functions. It will never be used by PCRE itself.
3753    
3754    (ii) The pcre_extra structure (see 45 above) has a void * field called
3755         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3756         pcre_callout_block structure has a field of the same name. The contents of
3757         the field passed in the pcre_extra structure are passed to the callout
3758         function in the corresponding field in the callout block. This makes it
3759         easier to use the same callout-containing regex from multiple threads. For
3760         testing, the pcretest program has a new data escape
3761    
3762           \C*n        pass the number n (may be negative) as callout_data
3763    
3764         If the callout function in pcretest receives a non-zero value as
3765         callout_data, it returns that value.
3766    
3767    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3768    there were some redundant $(CFLAGS) in commands that are now specified as
3769    $(LINK), which already includes $(CFLAGS).
3770    
3771    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3772    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3773    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3774    one-byte characters throughout. Note that case-insensitive matching applies
3775    only to characters whose values are less than 256. PCRE doesn't support the
3776    notion of cases for higher-valued characters.
3777    
3778    (i)   A character class whose characters are all within 0-255 is handled as
3779          a bit map, and the map is inverted for negative classes. Previously, a
3780          character > 255 always failed to match such a class; however it should
3781          match if the class was a negative one (e.g. [^ab]). This has been fixed.
3782    
3783    (ii)  A negated character class with a single character < 255 is coded as
3784          "not this character" (OP_NOT). This wasn't working properly when the test
3785          character was multibyte, either singly or repeated.
3786    
3787    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3788          mode, for example: \x{100}{2,3}.
3789    
3790    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3791          singly or repeated) now correctly test multibyte characters. However,
3792          PCRE doesn't recognize any characters with values greater than 255 as
3793          digits, spaces, or word characters. Such characters always match \D, \S,
3794          and \W, and never match \d, \s, or \w.
3795    
3796    (v)   Classes may now contain characters and character ranges with values
3797          greater than 255. For example: [ab\x{100}-\x{400}].
3798    
3799    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
3800          PCRE in UTF-8 mode.
3801    
3802    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3803    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3804    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3805    value.)
3806    
3807    53. The single man page has become too large. I have therefore split it up into
3808    a number of separate man pages. These also giv