/[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 754 by ph10, Sat Nov 19 18:32:18 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    
56    Version 8.20 21-Oct-2011
57    ------------------------
58    
59    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
60        a POSIX class. After further experiments with Perl, which convinced me that
61        Perl has bugs and confusions, a closing square bracket is no longer allowed
62        in a POSIX name. This bug also affected patterns with classes that started
63        with full stops.
64    
65    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
66        captured substring, but while checking the failing first alternative,
67        substring 1 is temporarily captured. If the output vector supplied to
68        pcre_exec() was not big enough for this capture, the yield of the function
69        was still zero ("insufficient space for captured substrings"). This cannot
70        be totally fixed without adding another stack variable, which seems a lot
71        of expense for a edge case. However, I have improved the situation in cases
72        such as /(a)(b)x|abc/ matched against "abc", where the return code
73        indicates that fewer than the maximum number of slots in the ovector have
74        been set.
75    
76    3.  Related to (2) above: when there are more back references in a pattern than
77        slots in the output vector, pcre_exec() uses temporary memory during
78        matching, and copies in the captures as far as possible afterwards. It was
79        using the entire output vector, but this conflicts with the specification
80        that only 2/3 is used for passing back captured substrings. Now it uses
81        only the first 2/3, for compatibility. This is, of course, another edge
82        case.
83    
84    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
85        main code base, and can be used by building with --enable-jit. When this is
86        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
87        runtime --no-jit option is given.
88    
89    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
90        ovector, the return from the function was zero, implying that there were
91        other matches that did not fit. The correct "exactly full" value is now
92        returned.
93    
94    6.  If a subpattern that was called recursively or as a subroutine contained
95        (*PRUNE) or any other control that caused it to give a non-standard return,
96        invalid errors such as "Error -26 (nested recursion at the same subject
97        position)" or even infinite loops could occur.
98    
99    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
100        computing the minimum length on reaching *ACCEPT, and so ended up with the
101        wrong value of 1 rather than 0. Further investigation indicates that
102        computing a minimum subject length in the presence of *ACCEPT is difficult
103        (think back references, subroutine calls), and so I have changed the code
104        so that no minimum is registered for a pattern that contains *ACCEPT.
105    
106    8.  If (*THEN) was present in the first (true) branch of a conditional group,
107        it was not handled as intended. [But see 16 below.]
108    
109    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
110        Sheri Pierce.
111    
112    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
113        the first byte in a match must be "a".
114    
115    11. Change 17 for 8.13 increased the recursion depth for patterns like
116        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
117        pattern contains any instances of (*THEN). If it does not, the old
118        optimizations are restored. It would be nice to do this on a per-group
119        basis, but at the moment that is not feasible.
120    
121    12. In some environments, the output of pcretest -C is CRLF terminated. This
122        broke RunTest's code that checks for the link size. A single white space
123        character after the value is now allowed for.
124    
125    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
126        For "fr", it uses the Windows-specific input and output files.
127    
128    14. If (*THEN) appeared in a group that was called recursively or as a
129        subroutine, it did not work as intended. [But see next item.]
130    
131    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
132        pattern fragments (but not containing any | characters). If A and B are
133        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
134        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
135        D. In other words, Perl considers parentheses that do not contain any |
136        characters to be part of a surrounding alternative, whereas PCRE was
137        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
138        differently. PCRE now behaves in the same way as Perl, except in the case
139        of subroutine/recursion calls such as (?1) which have in any case always
140        been different (but PCRE had them first :-).
141    
142    16. Related to 15 above: Perl does not treat the | in a conditional group as
143        creating alternatives. Such a group is treated in the same way as an
144        ordinary group without any | characters when processing (*THEN). PCRE has
145        been changed to match Perl's behaviour.
146    
147    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
148        RunGrepTest script failed.
149    
150    18. Change 22 for version 13 caused atomic groups to use more stack. This is
151        inevitable for groups that contain captures, but it can lead to a lot of
152        stack use in large patterns. The old behaviour has been restored for atomic
153        groups that do not contain any capturing parentheses.
154    
155    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
156        suppress the check for a minimum subject length at run time. (If it was
157        given to pcre_exec() or pcre_dfa_exec() it did work.)
158    
159    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
160        fail to work when decoding hex characters in data strings in EBCDIC
161        environments.
162    
163    21. It appears that in at least one Mac OS environment, the isxdigit() function
164        is implemented as a macro that evaluates to its argument more than once,
165        contravening the C 90 Standard (I haven't checked a later standard). There
166        was an instance in pcretest which caused it to go wrong when processing
167        \x{...} escapes in subject strings. The has been rewritten to avoid using
168        things like p++ in the argument of isxdigit().
169    
170    
171    Version 8.13 16-Aug-2011
172    ------------------------
173    
174    1.  The Unicode data tables have been updated to Unicode 6.0.0.
175    
176    2.  Two minor typos in pcre_internal.h have been fixed.
177    
178    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
179        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
180        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
181    
182    4.  There were a number of related bugs in the code for matching backrefences
183        caselessly in UTF-8 mode when codes for the characters concerned were
184        different numbers of bytes. For example, U+023A and U+2C65 are an upper
185        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
186        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
187        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
188        2-byte code at the end of the subject (it thought there wasn't enough data
189        left).
190    
191    5.  Comprehensive information about what went wrong is now returned by
192        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
193        as the output vector has at least 2 elements. The offset of the start of
194        the failing character and a reason code are placed in the vector.
195    
196    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
197        now returned is for the first byte of the failing character, instead of the
198        last byte inspected. This is an incompatible change, but I hope it is small
199        enough not to be a problem. It makes the returned offset consistent with
200        pcre_exec() and pcre_dfa_exec().
201    
202    7.  pcretest now gives a text phrase as well as the error number when
203        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
204        failure, the offset and reason code are output.
205    
206    8.  When \R was used with a maximizing quantifier it failed to skip backwards
207        over a \r\n pair if the subsequent match failed. Instead, it just skipped
208        back over a single character (\n). This seems wrong (because it treated the
209        two characters as a single entity when going forwards), conflicts with the
210        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
211        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
212        has been changed.
213    
214    9.  Some internal refactoring has changed the processing so that the handling
215        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
216        time (the PCRE_DOTALL option was changed this way some time ago: version
217        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
218        which was always a bit of a fudge. It also means that there is one less
219        argument for the match() function, which reduces its stack requirements
220        slightly. This change also fixes an incompatibility with Perl: the pattern
221        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
222    
223    10. More internal refactoring has drastically reduced the number of recursive
224        calls to match() for possessively repeated groups such as (abc)++ when
225        using pcre_exec().
226    
227    11. While implementing 10, a number of bugs in the handling of groups were
228        discovered and fixed:
229    
230        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
231        (a|)*(?1) gave a compile-time internal error.
232        ((a|)+)+  did not notice that the outer group could match an empty string.
233        (^a|^)+   was not marked as anchored.
234        (.*a|.*)+ was not marked as matching at start or after a newline.
235    
236    12. Yet more internal refactoring has removed another argument from the match()
237        function. Special calls to this function are now indicated by setting a
238        value in a variable in the "match data" data block.
239    
240    13. Be more explicit in pcre_study() instead of relying on "default" for
241        opcodes that mean there is no starting character; this means that when new
242        ones are added and accidentally left out of pcre_study(), testing should
243        pick them up.
244    
245    14. The -s option of pcretest has been documented for ages as being an old
246        synonym of -m (show memory usage). I have changed it to mean "force study
247        for every regex", that is, assume /S for every regex. This is similar to -i
248        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
249        using it. It makes it easier to run collections of tests with and without
250        study enabled, and thereby test pcre_study() more easily. All the standard
251        tests are now run with and without -s (but some patterns can be marked as
252        "never study" - see 20 below).
253    
254    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
255        restoration of the capturing data to the outer values was not happening
256        correctly.
257    
258    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
259        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
260        pattern had matched an empty string, and so incorrectly returned a no
261        match.
262    
263    17. There was optimizing code for the last branch of non-capturing parentheses,
264        and also for the obeyed branch of a conditional subexpression, which used
265        tail recursion to cut down on stack usage. Unfortunately, now that there is
266        the possibility of (*THEN) occurring in these branches, tail recursion is
267        no longer possible because the return has to be checked for (*THEN). These
268        two optimizations have therefore been removed. [But see 8.20/11 above.]
269    
270    18. If a pattern containing \R was studied, it was assumed that \R always
271        matched two bytes, thus causing the minimum subject length to be
272        incorrectly computed because \R can also match just one byte.
273    
274    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
275        was incorrectly computed.
276    
277    20. If /S is present twice on a test pattern in pcretest input, it now
278        *disables* studying, thereby overriding the use of -s on the command line
279        (see 14 above). This is necessary for one or two tests to keep the output
280        identical in both cases.
281    
282    21. When (*ACCEPT) was used in an assertion that matched an empty string and
283        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
284    
285    22. When an atomic group that contained a capturing parenthesis was
286        successfully matched, but the branch in which it appeared failed, the
287        capturing was not being forgotten if a higher numbered group was later
288        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
289        group 1 to "a", when in fact it should be unset. This applied to multi-
290        branched capturing and non-capturing groups, repeated or not, and also to
291        positive assertions (capturing in negative assertions does not happen
292        in PCRE) and also to nested atomic groups.
293    
294    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
295        subject after a captured substring, to make it easier to tell which of a
296        number of identical substrings has been captured.
297    
298    24. The way atomic groups are processed by pcre_exec() has been changed so that
299        if they are repeated, backtracking one repetition now resets captured
300        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
301        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
302        "aaa". Previously, it would have been "a". As part of this code
303        refactoring, the way recursive calls are handled has also been changed.
304    
305    25. If an assertion condition captured any substrings, they were not passed
306        back unless some other capturing happened later. For example, if
307        (?(?=(a))a) was matched against "a", no capturing was returned.
308    
309    26. When studying a pattern that contained subroutine calls or assertions,
310        the code for finding the minimum length of a possible match was handling
311        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
312        group 1 called group 2 while simultaneously a separate group 2 called group
313        1). A stack overflow occurred in this case. I have fixed this by limiting
314        the recursion depth to 10.
315    
316    27. Updated RunTest.bat in the distribution to the version supplied by Tom
317        Fortmann. This supports explicit test numbers on the command line, and has
318        argument validation and error reporting.
319    
320    28. An instance of \X with an unlimited repeat could fail if at any point the
321        first character it looked at was a mark character.
322    
323    29. Some minor code refactoring concerning Unicode properties and scripts
324        should reduce the stack requirement of match() slightly.
325    
326    30. Added the '=' option to pcretest to check the setting of unused capturing
327        slots at the end of the pattern, which are documented as being -1, but are
328        not included in the return count.
329    
330    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
331        compiled something random. Now it gives a compile-time error (as does
332        Perl).
333    
334    32. A *MARK encountered during the processing of a positive assertion is now
335        recorded and passed back (compatible with Perl).
336    
337    33. If --only-matching or --colour was set on a pcregrep call whose pattern
338        had alternative anchored branches, the search for a second match in a line
339        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
340        matched the line "0102" twice. The same bug affected patterns that started
341        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
342        twice.
343    
344    34. Previously, PCRE did not allow quantification of assertions. However, Perl
345        does, and because of capturing effects, quantifying parenthesized
346        assertions may at times be useful. Quantifiers are now allowed for
347        parenthesized assertions.
348    
349    35. A minor code tidy in pcre_compile() when checking options for \R usage.
350    
351    36. \g was being checked for fancy things in a character class, when it should
352        just be a literal "g".
353    
354    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
355        appearance of a nested POSIX class supersedes an apparent external class.
356        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
357        unescaped square brackets may also appear as part of class names. For
358        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
359        more like Perl. (But see 8.20/1 above.)
360    
361    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
362        was because it thought it was \N{name}, which is not supported).
363    
364    39. Add minix to OS list not supporting the -S option in pcretest.
365    
366    40. PCRE tries to detect cases of infinite recursion at compile time, but it
367        cannot analyze patterns in sufficient detail to catch mutual recursions
368        such as ((?1))((?2)). There is now a runtime test that gives an error if a
369        subgroup is called recursively as a subpattern for a second time at the
370        same position in the subject string. In previous releases this might have
371        been caught by the recursion limit, or it might have run out of stack.
372    
373    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
374        happen only once. PCRE was, however incorrectly giving a compile time error
375        "recursive call could loop indefinitely" because it cannot analyze the
376        pattern in sufficient detail. The compile time test no longer happens when
377        PCRE is compiling a conditional subpattern, but actual runaway loops are
378        now caught at runtime (see 40 above).
379    
380    42. It seems that Perl allows any characters other than a closing parenthesis
381        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
382        has been changed to be the same.
383    
384    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
385        as not to get warnings when autogen.sh is called. Also changed
386        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
387    
388    44. To help people who use pcregrep to scan files containing exceedingly long
389        lines, the following changes have been made:
390    
391        (a) The default value of the buffer size parameter has been increased from
392            8K to 20K. (The actual buffer used is three times this size.)
393    
394        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
395            PCRE is built.
396    
397        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
398            to be set at run time.
399    
400        (d) Numerical values in pcregrep options can be followed by K or M, for
401            example --buffer-size=50K.
402    
403        (e) If a line being scanned overflows pcregrep's buffer, an error is now
404            given and the return code is set to 2.
405    
406    45. Add a pointer to the latest mark to the callout data block.
407    
408    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
409        partial match of an empty string instead of no match. This was specific to
410        the use of ".".
411    
412    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
413        complete match instead of a partial match. This bug was dependent on both
414        the PCRE_UTF8 and PCRE_DOTALL options being set.
415    
416    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
417        starting byte set, because \b was not being ignored.
418    
419    
420    Version 8.12 15-Jan-2011
421    ------------------------
422    
423    1.  Fixed some typos in the markup of the man pages, and wrote a script that
424        checks for such things as part of the documentation building process.
425    
426    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
427        --match-limit and --recursion-limit options (added for 8.11). In
428        particular, this made one of the standard tests fail. (The integer value
429        went into the wrong half of a long int.)
430    
431    3.  If the --colour option was given to pcregrep with -v (invert match), it
432        did strange things, either producing crazy output, or crashing. It should,
433        of course, ignore a request for colour when reporting lines that do not
434        match.
435    
436    4.  Another pcregrep bug caused similar problems if --colour was specified with
437        -M (multiline) and the pattern match finished with a line ending.
438    
439    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
440        matched in multiline mode, the following line was shown as part of the
441        match. This seems wrong, so I have changed it.
442    
443    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
444        the check for further matches in the same line (so they could be coloured)
445        to overrun the end of the current line. If another match was found, it was
446        incorrectly shown (and then shown again when found in the next line).
447    
448    7.  If pcregrep was compiled under Windows, there was a reference to the
449        function pcregrep_exit() before it was defined. I am assuming this was
450        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
451        reported by a user. I've moved the definition above the reference.
452    
453    
454    Version 8.11 10-Dec-2010
455    ------------------------
456    
457    1.  (*THEN) was not working properly if there were untried alternatives prior
458        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
459        backtracked to try for "b" instead of moving to the next alternative branch
460        at the same level (in this case, to look for "c"). The Perl documentation
461        is clear that when (*THEN) is backtracked onto, it goes to the "next
462        alternative in the innermost enclosing group".
463    
464    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
465        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
466        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
467        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
468        (*THEN).
469    
470    3.  If \s appeared in a character class, it removed the VT character from
471        the class, even if it had been included by some previous item, for example
472        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
473        of \s, but is part of the POSIX "space" class.)
474    
475    4.  A partial match never returns an empty string (because you can always
476        match an empty string at the end of the subject); however the checking for
477        an empty string was starting at the "start of match" point. This has been
478        changed to the "earliest inspected character" point, because the returned
479        data for a partial match starts at this character. This means that, for
480        example, /(?<=abc)def/ gives a partial match for the subject "abc"
481        (previously it gave "no match").
482    
483    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
484        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
485        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
486        has an implication that the given string is incomplete (because a partial
487        match is preferred over a full match). For this reason, these items now
488        give a partial match in this situation. [Aside: previously, the one case
489        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
490        match rather than a full match, which was wrong by the old rules, but is
491        now correct.]
492    
493    6.  There was a bug in the handling of #-introduced comments, recognized when
494        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
495        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
496        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
497        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
498        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
499        places in pcre_compile().
500    
501    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
502        comments when looking ahead for named forward references to subpatterns,
503        the only newline sequence it recognized was NL. It now handles newlines
504        according to the set newline convention.
505    
506    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
507        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
508        cater for a lack of strerror(). These oversights have been fixed.
509    
510    9.  Added --match-limit and --recursion-limit to pcregrep.
511    
512    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
513    
514    11. When the -o option was used, pcregrep was setting a return code of 1, even
515        when matches were found, and --line-buffered was not being honoured.
516    
517    12. Added an optional parentheses number to the -o and --only-matching options
518        of pcregrep.
519    
520    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
521        can match an empty string. The code to do it in pcretest and pcredemo
522        needed fixing:
523    
524        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
525            only one byte after an empty string match just before CRLF (this case
526            just got forgotten; "any" and "anycrlf" were OK).
527    
528        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
529            mode when an empty string match preceded an ASCII character followed by
530            a non-ASCII character. (The code for advancing by one character rather
531            than one byte was nonsense.)
532    
533        (c) The pcredemo.c sample program did not have any code at all to handle
534            the cases when CRLF is a valid newline sequence.
535    
536    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
537        as a starting offset was within the subject string. There is now a new
538        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
539        negative or greater than the length of the string. In order to test this,
540        pcretest is extended to allow the setting of negative starting offsets.
541    
542    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
543        starting offset points to the beginning of a UTF-8 character was
544        unnecessarily clumsy. I tidied it up.
545    
546    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
547        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
548    
549    17. Nobody had reported that the --include_dir option, which was added in
550        release 7.7 should have been called --include-dir (hyphen, not underscore)
551        for compatibility with GNU grep. I have changed it to --include-dir, but
552        left --include_dir as an undocumented synonym, and the same for
553        --exclude-dir, though that is not available in GNU grep, at least as of
554        release 2.5.4.
555    
556    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
557        characters from a string of bytes) have been redefined so as not to use
558        loops, in order to improve performance in some environments. At the same
559        time, I abstracted some of the common code into auxiliary macros to save
560        repetition (this should not affect the compiled code).
561    
562    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
563        compile-time error is now given if \c is not followed by an ASCII
564        character, that is, a byte less than 128. (In EBCDIC mode, the code is
565        different, and any byte value is allowed.)
566    
567    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
568        START_OPTIMIZE option, which is now allowed at compile time - but just
569        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
570        to pcregrep and other applications that have no direct access to PCRE
571        options. The new /Y option in pcretest sets this option when calling
572        pcre_compile().
573    
574    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
575        back references. Groups containing recursive back references were forced to
576        be atomic by that change, but in the case of named groups, the amount of
577        memory required was incorrectly computed, leading to "Failed: internal
578        error: code overflow". This has been fixed.
579    
580    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
581        pcretest.c, to avoid build problems in some Borland environments.
582    
583    
584    Version 8.10 25-Jun-2010
585    ------------------------
586    
587    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
588        THEN.
589    
590    2.  (*ACCEPT) was not working when inside an atomic group.
591    
592    3.  Inside a character class, \B is treated as a literal by default, but
593        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
594        causes the error). The code is unchanged, but I tidied the documentation.
595    
596    4.  Inside a character class, PCRE always treated \R and \X as literals,
597        whereas Perl faults them if its -w option is set. I have changed PCRE so
598        that it faults them when PCRE_EXTRA is set.
599    
600    5.  Added support for \N, which always matches any character other than
601        newline. (It is the same as "." when PCRE_DOTALL is not set.)
602    
603    6.  When compiling pcregrep with newer versions of gcc which may have
604        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
605        declared with attribute warn_unused_result" were given. Just casting the
606        result to (void) does not stop the warnings; a more elaborate fudge is
607        needed. I've used a macro to implement this.
608    
609    7.  Minor change to pcretest.c to avoid a compiler warning.
610    
611    8.  Added four artifical Unicode properties to help with an option to make
612        \s etc use properties (see next item). The new properties are: Xan
613        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
614    
615    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
616        use Unicode properties. (*UCP) at the start of a pattern can be used to set
617        this option. Modified pcretest to add /W to test this facility. Added
618        REG_UCP to make it available via the POSIX interface.
619    
620    10. Added --line-buffered to pcregrep.
621    
622    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
623        studied, and the match started with a letter with a code point greater than
624        127 whose first byte was different to the first byte of the other case of
625        the letter, the other case of this starting letter was not recognized
626        (#976).
627    
628    12. If a pattern that was studied started with a repeated Unicode property
629        test, for example, \p{Nd}+, there was the theoretical possibility of
630        setting up an incorrect bitmap of starting bytes, but fortunately it could
631        not have actually happened in practice until change 8 above was made (it
632        added property types that matched character-matching opcodes).
633    
634    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
635        possible starting bytes for non-anchored patterns.
636    
637    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
638        \R, and also a number of cases that involve Unicode properties, both
639        explicit and implicit when PCRE_UCP is set.
640    
641    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
642        input, it could crash or give wrong results if characters with values
643        greater than 0xc0 were present in the subject string. (Detail: it assumed
644        UTF-8 input when processing these items.)
645    
646    16. Added a lot of (int) casts to avoid compiler warnings in systems where
647        size_t is 64-bit (#991).
648    
649    17. Added a check for running out of memory when PCRE is compiled with
650        --disable-stack-for-recursion (#990).
651    
652    18. If the last data line in a file for pcretest does not have a newline on
653        the end, a newline was missing in the output.
654    
655    19. The default pcre_chartables.c file recognizes only ASCII characters (values
656        less than 128) in its various bitmaps. However, there is a facility for
657        generating tables according to the current locale when PCRE is compiled. It
658        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
659        characters, are recognized by isspace() and therefore were getting set in
660        these tables, and indeed these tables seem to approximate to ISO 8859. This
661        caused a problem in UTF-8 mode when pcre_study() was used to create a list
662        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
663        which of course cannot start UTF-8 characters. I have changed the code so
664        that only real ASCII characters (less than 128) and the correct starting
665        bytes for UTF-8 encodings are set for characters greater than 127 when in
666        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
667        altogether.)
668    
669    20. Added the /T option to pcretest so as to be able to run tests with non-
670        standard character tables, thus making it possible to include the tests
671        used for 19 above in the standard set of tests.
672    
673    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
674        reference to a subpattern the other side of a comment that contains an
675        opening parenthesis caused either an internal compiling error, or a
676        reference to the wrong subpattern.
677    
678    
679    Version 8.02 19-Mar-2010
680    ------------------------
681    
682    1.  The Unicode data tables have been updated to Unicode 5.2.0.
683    
684    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
685        configured.
686    
687    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
688        original author of that file, following a query about its status.
689    
690    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
691        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
692    
693    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
694        quantifier applied to a forward-referencing subroutine call, could compile
695        incorrect code or give the error "internal error: previously-checked
696        referenced subpattern not found".
697    
698    6.  Both MS Visual Studio and Symbian OS have problems with initializing
699        variables to point to external functions. For these systems, therefore,
700        pcre_malloc etc. are now initialized to local functions that call the
701        relevant global functions.
702    
703    7.  There were two entries missing in the vectors called coptable and poptable
704        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
705        I've fixed the data, and added a kludgy way of testing at compile time that
706        the lengths are correct (equal to the number of opcodes).
707    
708    8.  Following on from 7, I added a similar kludge to check the length of the
709        eint vector in pcreposix.c.
710    
711    9.  Error texts for pcre_compile() are held as one long string to avoid too
712        much relocation at load time. To find a text, the string is searched,
713        counting zeros. There was no check for running off the end of the string,
714        which could happen if a new error number was added without updating the
715        string.
716    
717    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
718    
719    11. \K was not working if it appeared in an atomic group or in a group that
720        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
721        \K is "not well defined" if used in an assertion. PCRE now accepts it if
722        the assertion is positive, but not if it is negative.
723    
724    12. Change 11 fortuitously reduced the size of the stack frame used in the
725        "match()" function of pcre_exec.c by one pointer. Forthcoming
726        implementation of support for (*MARK) will need an extra pointer on the
727        stack; I have reserved it now, so that the stack frame size does not
728        decrease.
729    
730    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
731        item in branch that calls a recursion is a subroutine call - as in the
732        second branch in the above example - was incorrectly given the compile-
733        time error "recursive call could loop indefinitely" because pcre_compile()
734        was not correctly checking the subroutine for matching a non-empty string.
735    
736    14. The checks for overrunning compiling workspace could trigger after an
737        overrun had occurred. This is a "should never occur" error, but it can be
738        triggered by pathological patterns such as hundreds of nested parentheses.
739        The checks now trigger 100 bytes before the end of the workspace.
740    
741    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
742    
743    
744    Version 8.01 19-Jan-2010
745    ------------------------
746    
747    1.  If a pattern contained a conditional subpattern with only one branch (in
748        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
749        computed the wrong minimum data length (which is of course zero for such
750        subpatterns). This could cause incorrect "no match" results.
751    
752    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
753        the pattern is reset in the first branch, pcre_compile() failed with
754        "internal error: code overflow at offset...". This happened only when
755        the reset was to the original external option setting. (An optimization
756        abstracts leading options settings into an external setting, which was the
757        cause of this.)
758    
759    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
760        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
761        assertion pattern did not match (meaning that the assertion was true), it
762        was incorrectly treated as false if the SKIP had been reached during the
763        matching. This also applied to assertions used as conditions.
764    
765    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
766        assertion subpattern, including such a pattern used as a condition,
767        unpredictable results occurred, instead of the error return
768        PCRE_ERROR_DFA_UITEM.
769    
770    5.  The C++ GlobalReplace function was not working like Perl for the special
771        situation when an empty string is matched. It now does the fancy magic
772        stuff that is necessary.
773    
774    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
775        removed. (These were left over from very, very early versions of PCRE.)
776    
777    7.  Some cosmetic changes to the code to make life easier when compiling it
778        as part of something else:
779    
780        (a) Change DEBUG to PCRE_DEBUG.
781    
782        (b) In pcre_compile(), rename the member of the "branch_chain" structure
783            called "current" as "current_branch", to prevent a collision with the
784            Linux macro when compiled as a kernel module.
785    
786        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
787            prevent a collision with the Linux macro when compiled as a kernel
788            module.
789    
790    8.  In pcre_compile() there are some checks for integer overflows that used to
791        cast potentially large values to (double). This has been changed to that
792        when building, a check for int64_t is made, and if it is found, it is used
793        instead, thus avoiding the use of floating point arithmetic. (There is no
794        other use of FP in PCRE.) If int64_t is not found, the fallback is to
795        double.
796    
797    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
798        2005 (difference between two addresses compared to an unsigned value).
799    
800    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
801        custom one, because of the following reported problem in Windows:
802    
803          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
804              under Win32.
805          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
806              therefore missing the function definition.
807          - The compiler thus generates a "C" signature for the test function.
808          - The linker fails to find the "C" function.
809          - PCRE fails to configure if asked to do so against libbz2.
810    
811    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
812        messages were output:
813    
814          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
815          rerunning libtoolize, to keep the correct libtool macros in-tree.
816          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
817    
818        I have done both of these things.
819    
820    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
821        most of the time, it *can* run out if it is given a pattern that contains a
822        runaway infinite recursion. I updated the discussion in the pcrestack man
823        page.
824    
825    13. Now that we have gone to the x.xx style of version numbers, the minor
826        version may start with zero. Using 08 or 09 is a bad idea because users
827        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
828        interpreted as invalid octal numbers. I've updated the previous comment in
829        configure.ac, and also added a check that gives an error if 08 or 09 are
830        used.
831    
832    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
833        causing partial matching to fail when the end of the subject matched \W
834        in a UTF-8 pattern where \W was quantified with a minimum of 3.
835    
836    15. There were some discrepancies between the declarations in pcre_internal.h
837        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
838        their definitions. The declarations used "const uschar *" and the
839        definitions used USPTR. Even though USPTR is normally defined as "const
840        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
841        reported that: "This difference in casting confuses some C++ compilers, for
842        example, SunCC recognizes above declarations as different functions and
843        generates broken code for hbpcre." I have changed the declarations to use
844        USPTR.
845    
846    16. GNU libtool is named differently on some systems. The autogen.sh script now
847        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
848        (FreeBSD).
849    
850    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
851        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
852        comment: "Figure out how to create a longlong from a string: strtoll and
853        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
854        instance, but it only takes 2 args instead of 3!"
855    
856    18. A subtle bug concerned with back references has been fixed by a change of
857        specification, with a corresponding code fix. A pattern such as
858        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
859        refers, was giving matches when it shouldn't. For example, xa=xaaa would
860        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
861        same bug. Such groups have to be quantified to be useful, or contained
862        inside another quantified group. (If there's no repetition, the reference
863        can never match.) The problem arises because, having left the group and
864        moved on to the rest of the pattern, a later failure that backtracks into
865        the group uses the captured value from the final iteration of the group
866        rather than the correct earlier one. I have fixed this in PCRE by forcing
867        any group that contains a reference to itself to be an atomic group; that
868        is, there cannot be any backtracking into it once it has completed. This is
869        similar to recursive and subroutine calls.
870    
871    
872    Version 8.00 19-Oct-09
873    ----------------------
874    
875    1.  The table for translating pcre_compile() error codes into POSIX error codes
876        was out-of-date, and there was no check on the pcre_compile() error code
877        being within the table. This could lead to an OK return being given in
878        error.
879    
880    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
881        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
882        in a Windows environment.
883    
884    3.  The pcregrep --count option prints the count for each file even when it is
885        zero, as does GNU grep. However, pcregrep was also printing all files when
886        --files-with-matches was added. Now, when both options are given, it prints
887        counts only for those files that have at least one match. (GNU grep just
888        prints the file name in this circumstance, but including the count seems
889        more useful - otherwise, why use --count?) Also ensured that the
890        combination -clh just lists non-zero counts, with no names.
891    
892    4.  The long form of the pcregrep -F option was incorrectly implemented as
893        --fixed_strings instead of --fixed-strings. This is an incompatible change,
894        but it seems right to fix it, and I didn't think it was worth preserving
895        the old behaviour.
896    
897    5.  The command line items --regex=pattern and --regexp=pattern were not
898        recognized by pcregrep, which required --regex pattern or --regexp pattern
899        (with a space rather than an '='). The man page documented the '=' forms,
900        which are compatible with GNU grep; these now work.
901    
902    6.  No libpcreposix.pc file was created for pkg-config; there was just
903        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
904    
905    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
906        when UCP support is not needed, by modifying the Python script that
907        generates it from Unicode data files. This should not matter if the module
908        is correctly used as a library, but I received one complaint about 50K of
909        unwanted data. My guess is that the person linked everything into his
910        program rather than using a library. Anyway, it does no harm.
911    
912    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
913        was a minimum greater than 1 for a wide character in a possessive
914        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
915        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
916        character. Chaos in the form of incorrect output or a compiling loop could
917        result.
918    
919    9.  The restrictions on what a pattern can contain when partial matching is
920        requested for pcre_exec() have been removed. All patterns can now be
921        partially matched by this function. In addition, if there are at least two
922        slots in the offset vector, the offset of the earliest inspected character
923        for the match and the offset of the end of the subject are set in them when
924        PCRE_ERROR_PARTIAL is returned.
925    
926    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
927        synonymous with PCRE_PARTIAL, for backwards compatibility, and
928        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
929        and may be more useful for multi-segment matching.
930    
931    11. Partial matching with pcre_exec() is now more intuitive. A partial match
932        used to be given if ever the end of the subject was reached; now it is
933        given only if matching could not proceed because another character was
934        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
935        string "Z", which now yields "no match" instead of "partial match". In the
936        case of pcre_dfa_exec(), "no match" is given if every matching path for the
937        final character ended with (*FAIL).
938    
939    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
940        if the pattern had a "must contain" character that was already found in the
941        earlier partial match, unless partial matching was again requested. For
942        example, with the pattern /dog.(body)?/, the "must contain" character is
943        "g". If the first part-match was for the string "dog", restarting with
944        "sbody" failed. This bug has been fixed.
945    
946    13. The string returned by pcre_dfa_exec() after a partial match has been
947        changed so that it starts at the first inspected character rather than the
948        first character of the match. This makes a difference only if the pattern
949        starts with a lookbehind assertion or \b or \B (\K is not supported by
950        pcre_dfa_exec()). It's an incompatible change, but it makes the two
951        matching functions compatible, and I think it's the right thing to do.
952    
953    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
954        so that the demonstration program is easily available in environments where
955        PCRE has not been installed from source.
956    
957    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
958        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
959        library.
960    
961    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
962        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
963        is not the first non-POSIX option to be added. Clearly some people find
964        these options useful.
965    
966    17. If a caller to the POSIX matching function regexec() passes a non-zero
967        value for nmatch with a NULL value for pmatch, the value of
968        nmatch is forced to zero.
969    
970    18. RunGrepTest did not have a test for the availability of the -u option of
971        the diff command, as RunTest does. It now checks in the same way as
972        RunTest, and also checks for the -b option.
973    
974    19. If an odd number of negated classes containing just a single character
975        interposed, within parentheses, between a forward reference to a named
976        subpattern and the definition of the subpattern, compilation crashed with
977        an internal error, complaining that it could not find the referenced
978        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
979        [The bug was that it was starting one character too far in when skipping
980        over the character class, thus treating the ] as data rather than
981        terminating the class. This meant it could skip too much.]
982    
983    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
984        /g option in pcretest when the pattern contains \K, which makes it possible
985        to have an empty string match not at the start, even when the pattern is
986        anchored. Updated pcretest and pcredemo to use this option.
987    
988    21. If the maximum number of capturing subpatterns in a recursion was greater
989        than the maximum at the outer level, the higher number was returned, but
990        with unset values at the outer level. The correct (outer level) value is
991        now given.
992    
993    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
994        PCRE did not set those parentheses (unlike Perl). I have now found a way to
995        make it do so. The string so far is captured, making this feature
996        compatible with Perl.
997    
998    23. The tests have been re-organized, adding tests 11 and 12, to make it
999        possible to check the Perl 5.10 features against Perl 5.10.
1000    
1001    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1002        pattern matches a fixed length string. PCRE did not allow this; now it
1003        does. Neither allows recursion.
1004    
1005    25. I finally figured out how to implement a request to provide the minimum
1006        length of subject string that was needed in order to match a given pattern.
1007        (It was back references and recursion that I had previously got hung up
1008        on.) This code has now been added to pcre_study(); it finds a lower bound
1009        to the length of subject needed. It is not necessarily the greatest lower
1010        bound, but using it to avoid searching strings that are too short does give
1011        some useful speed-ups. The value is available to calling programs via
1012        pcre_fullinfo().
1013    
1014    26. While implementing 25, I discovered to my embarrassment that pcretest had
1015        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1016        study optimizations had never been tested with that matching function.
1017        Oops. What is worse, even when it was passed study data, there was a bug in
1018        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1019        were also very few tests of studied patterns with pcre_dfa_exec().
1020    
1021    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1022        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1023        on the other side of the coin, they are no longer allowed to have different
1024        names, because these cannot be distinguished in PCRE, and this has caused
1025        confusion. (This is a difference from Perl.)
1026    
1027    28. When duplicate subpattern names are present (necessarily with different
1028        numbers, as required by 27 above), and a test is made by name in a
1029        conditional pattern, either for a subpattern having been matched, or for
1030        recursion in such a pattern, all the associated numbered subpatterns are
1031        tested, and the overall condition is true if the condition is true for any
1032        one of them. This is the way Perl works, and is also more like the way
1033        testing by number works.
1034    
1035    
1036    Version 7.9 11-Apr-09
1037    ---------------------
1038    
1039    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1040        (pcretest), all targets were linked against these libraries. This included
1041        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1042        libraries. This caused unwanted dependencies to be created. This problem
1043        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1044        pcretest is linked with readline.
1045    
1046    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1047        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1048        moved outside it again, because FALSE and TRUE are already defined in AIX,
1049        but BOOL is not.
1050    
1051    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1052        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1053    
1054    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1055        hyphens) following file names and line numbers when outputting matching
1056        lines. This is not true; no spaces are inserted. I have also clarified the
1057        wording for the --colour (or --color) option.
1058    
1059    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1060        was not coloured; this is different to GNU grep, so I have changed it to be
1061        the same.
1062    
1063    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1064        each matching line was coloured. Now it goes on to look for further matches
1065        of any of the test patterns, which is the same behaviour as GNU grep.
1066    
1067    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1068        doesn't make sense to accept an empty string match in pcregrep, so I have
1069        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1070        seems to be how GNU grep behaves.
1071    
1072    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1073        start or after a newline", because the conditional assertion was not being
1074        correctly handled. The rule now is that both the assertion and what follows
1075        in the first alternative must satisfy the test.
1076    
1077    9.  If auto-callout was enabled in a pattern with a conditional group whose
1078        condition was an assertion, PCRE could crash during matching, both with
1079        pcre_exec() and pcre_dfa_exec().
1080    
1081    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1082        used for matching.
1083    
1084    11. Unicode property support in character classes was not working for
1085        characters (bytes) greater than 127 when not in UTF-8 mode.
1086    
1087    12. Added the -M command line option to pcretest.
1088    
1089    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1090    
1091    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1092    
1093    16. Added comments and documentation about mis-use of no_arg in the C++
1094        wrapper.
1095    
1096    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1097        from Martin Jerabek that uses macro names for all relevant character and
1098        string constants.
1099    
1100    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1101        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1102        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1103        these, but not everybody uses configure.
1104    
1105    19. A conditional group that had only one branch was not being correctly
1106        recognized as an item that could match an empty string. This meant that an
1107        enclosing group might also not be so recognized, causing infinite looping
1108        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1109        with the subject "ab", where knowledge that the repeated group can match
1110        nothing is needed in order to break the loop.
1111    
1112    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1113        exec(), but without supplying a callout function, matching went wrong.
1114    
1115    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1116        leak if the size of the offset vector was greater than 30. When the vector
1117        is smaller, the saved offsets during recursion go onto a local stack
1118        vector, but for larger vectors malloc() is used. It was failing to free
1119        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1120        error, in fact).
1121    
1122    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1123        heapframe that is used only when UTF-8 support is enabled. This caused no
1124        problem, but was untidy.
1125    
1126    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1127        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1128        included within another project.
1129    
1130    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1131        slightly modified by me:
1132    
1133          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1134              not building pcregrep.
1135    
1136          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1137              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1138    
1139    25. Forward references, both numeric and by name, in patterns that made use of
1140        duplicate group numbers, could behave incorrectly or give incorrect errors,
1141        because when scanning forward to find the reference group, PCRE was not
1142        taking into account the duplicate group numbers. A pattern such as
1143        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1144    
1145    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1146        the feature of a custom pointer more persuasive (as requested by a user).
1147    
1148    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1149        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1150        pre-defined.
1151    
1152    28. Added support for (*UTF8) at the start of a pattern.
1153    
1154    29. Arrange for flags added by the "release type" setting in CMake to be shown
1155        in the configuration summary.
1156    
1157    
1158    Version 7.8 05-Sep-08
1159    ---------------------
1160    
1161    1.  Replaced UCP searching code with optimized version as implemented for Ad
1162        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1163        stage table and inline lookup instead of a function, giving speed ups of 2
1164        to 5 times on some simple patterns that I tested. Permission was given to
1165        distribute the MultiStage2.py script that generates the tables (it's not in
1166        the tarball, but is in the Subversion repository).
1167    
1168    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1169        scripts.
1170    
1171    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1172        a group with a zero qualifier. The result of the study could be incorrect,
1173        or the function might crash, depending on the pattern.
1174    
1175    4.  Caseless matching was not working for non-ASCII characters in back
1176        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1177        It now works when Unicode Property Support is available.
1178    
1179    5.  In pcretest, an escape such as \x{de} in the data was always generating
1180        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1181        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1182        truncation.
1183    
1184    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1185    
1186    7.  Added two (int) casts to pcregrep when printing the difference of two
1187        pointers, in case they are 64-bit values.
1188    
1189    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1190        test 2 if it fails.
1191    
1192    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1193        and a #define of that name to empty if it is not externally set. This is to
1194        allow users of MSVC to set it if necessary.
1195    
1196    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1197        the convenience functions in the pcre_get.c source file.
1198    
1199    11. An option change at the start of a pattern that had top-level alternatives
1200        could cause overwriting and/or a crash. This command provoked a crash in
1201        some environments:
1202    
1203          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1204    
1205        This potential security problem was recorded as CVE-2008-2371.
1206    
1207    12. For a pattern where the match had to start at the beginning or immediately
1208        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1209        pcre_dfa_exec() could read past the end of the passed subject if there was
1210        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1211        pcretest so that it places the subject at the end of its malloc-ed buffer.
1212    
1213    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1214        exec() might read past the end of the data buffer in UTF-8 mode.
1215    
1216    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1217        the data contained the byte 0x85 as part of a UTF-8 character within its
1218        first line. This applied both to normal and DFA matching.
1219    
1220    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1221        /^[^d]*?$/8 failed to match "abc".
1222    
1223    16. Added a missing copyright notice to pcrecpp_internal.h.
1224    
1225    17. Make it more clear in the documentation that values returned from
1226        pcre_exec() in ovector are byte offsets, not character counts.
1227    
1228    18. Tidied a few places to stop certain compilers from issuing warnings.
1229    
1230    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1231        supplied by Stefan Weber. I made a further small update for 7.8 because
1232        there is a change of source arrangements: the pcre_searchfuncs.c module is
1233        replaced by pcre_ucd.c.
1234    
1235    
1236    Version 7.7 07-May-08
1237    ---------------------
1238    
1239    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1240        a string to a long long, pretend we don't even have a long long." This is
1241        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1242    
1243    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1244        pre-7.6 versions, which defined a global no_arg variable instead of putting
1245        it in the RE class. (See also #8 below.)
1246    
1247    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1248        Lopes.
1249    
1250    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1251    
1252        (1) The include/exclude patterns were being applied to the whole pathnames
1253            of files, instead of just to the final components.
1254    
1255        (2) If there was more than one level of directory, the subdirectories were
1256            skipped unless they satisfied the include/exclude conditions. This is
1257            inconsistent with GNU grep (and could even be seen as contrary to the
1258            pcregrep specification - which I improved to make it absolutely clear).
1259            The action now is always to scan all levels of directory, and just
1260            apply the include/exclude patterns to regular files.
1261    
1262    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1263        --exclude_dir in the tests to avoid scanning .svn directories.
1264    
1265    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1266        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1267        doesn't support NULs in patterns.
1268    
1269    7.  Added some missing "const"s to declarations of static tables in
1270        pcre_compile.c and pcre_dfa_exec.c.
1271    
1272    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1273        caused by fix #2  above. (Subsequently also a second patch to fix the
1274        first patch. And a third patch - this was a messy problem.)
1275    
1276    9.  Applied Craig's patch to remove the use of push_back().
1277    
1278    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1279        matching function regexec().
1280    
1281    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1282        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1283        references. PCRE supports relative numbers with this syntax (I don't think
1284        Oniguruma does).
1285    
1286    12. Previously, a group with a zero repeat such as (...){0} was completely
1287        omitted from the compiled regex. However, this means that if the group
1288        was called as a subroutine from elsewhere in the pattern, things went wrong
1289        (an internal error was given). Such groups are now left in the compiled
1290        pattern, with a new opcode that causes them to be skipped at execution
1291        time.
1292    
1293    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1294        to the way PCRE behaves:
1295    
1296        (a) A lone ] character is dis-allowed (Perl treats it as data).
1297    
1298        (b) A back reference to an unmatched subpattern matches an empty string
1299            (Perl fails the current match path).
1300    
1301        (c) A data ] in a character class must be notated as \] because if the
1302            first data character in a class is ], it defines an empty class. (In
1303            Perl it is not possible to have an empty class.) The empty class []
1304            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1305            The negative empty class [^] matches any one character, independently
1306            of the DOTALL setting.
1307    
1308    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1309        non-existent subpattern following a character class starting with ']' and
1310        containing () gave an internal compiling error instead of "reference to
1311        non-existent subpattern". Fortunately, when the pattern did exist, the
1312        compiled code was correct. (When scanning forwards to check for the
1313        existencd of the subpattern, it was treating the data ']' as terminating
1314        the class, so got the count wrong. When actually compiling, the reference
1315        was subsequently set up correctly.)
1316    
1317    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1318        it was being rejected as not supported by pcre_dfa_exec(), even though
1319        other assertions are supported. I have made pcre_dfa_exec() support
1320        (*FAIL).
1321    
1322    16. The implementation of 13c above involved the invention of a new opcode,
1323        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1324        cannot be changed at match time, I realized I could make a small
1325        improvement to matching performance by compiling OP_ALLANY instead of
1326        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1327        on the OP_ANY path.
1328    
1329    17. Compiling pcretest on Windows with readline support failed without the
1330        following two fixes: (1) Make the unistd.h include conditional on
1331        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1332    
1333    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1334        ncurses library to be included for pcretest when ReadLine support is
1335        requested, but also to allow for it to be overridden. This patch came from
1336        Daniel Bergström.
1337    
1338    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1339        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1340        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1341        spotting this.
1342    
1343    
1344    Version 7.6 28-Jan-08
1345    ---------------------
1346    
1347    1.  A character class containing a very large number of characters with
1348        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1349        overflow.
1350    
1351    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1352        HAVE_LONG_LONG is not defined.
1353    
1354    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1355        bring them up to date and include new features. This patch includes:
1356    
1357        - Fixed PH's badly added libz and libbz2 support.
1358        - Fixed a problem with static linking.
1359        - Added pcredemo. [But later removed - see 7 below.]
1360        - Fixed dftables problem and added an option.
1361        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1362            HAVE_LONG_LONG.
1363        - Added readline support for pcretest.
1364        - Added an listing of the option settings after cmake has run.
1365    
1366    4.  A user submitted a patch to Makefile that makes it easy to create
1367        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1368        Makefile.am that cause it to include this special target, without
1369        affecting anything else. Note that the same mingw target plus all
1370        the other distribution libraries and programs are now supported
1371        when configuring with CMake (see 6 below) instead of with
1372        Configure/Make.
1373    
1374    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1375        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1376        exported in the Windows port". It has not yet been confirmed that the patch
1377        solves the problem, but it does no harm.
1378    
1379    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1380        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1381        with CMake, and also correct the comment about stack recursion.
1382    
1383    7.  Remove the automatic building of pcredemo from the ./configure system and
1384        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1385        of a program that users should build themselves after PCRE is installed, so
1386        building it automatically is not really right. What is more, it gave
1387        trouble in some build environments.
1388    
1389    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1390    
1391    
1392    Version 7.5 10-Jan-08
1393    ---------------------
1394    
1395    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1396        values in parens when parsing an RE using the C++ wrapper."
1397    
1398    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1399        Characters greater than 255 were excluded from the class instead of being
1400        included.
1401    
1402    3.  The same bug as (2) above applied to negated POSIX classes such as
1403        [:^space:].
1404    
1405    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1406        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1407        I have changed it.
1408    
1409    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1410        first named subpattern) and a construct such as (?&a) would reference the
1411        first named subpattern whose name started with "a" (in other words, the
1412        length check was missing). Both these problems are fixed. "Subpattern name
1413        expected" is now given for (?&) (a zero-length name), and this patch also
1414        makes it give the same error for \k'' (previously it complained that that
1415        was a reference to a non-existent subpattern).
1416    
1417    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1418        this is right because (?- can be followed by option settings as well as by
1419        digits. I have, however, made the messages clearer.
1420    
1421    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1422        than the number used in the conditional) now cause a compile-time error.
1423        This is actually not compatible with Perl, which accepts such patterns, but
1424        treats the conditional as always being FALSE (as PCRE used to), but it
1425        seems to me that giving a diagnostic is better.
1426    
1427    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1428        and messages.
1429    
1430    9.  Fix two occurrences of "backslash" in comments that should have been
1431        "backspace".
1432    
1433    10. Remove two redundant lines of code that can never be obeyed (their function
1434        was moved elsewhere).
1435    
1436    11. The program that makes PCRE's Unicode character property table had a bug
1437        which caused it to generate incorrect table entries for sequences of
1438        characters that have the same character type, but are in different scripts.
1439        It amalgamated them into a single range, with the script of the first of
1440        them. In other words, some characters were in the wrong script. There were
1441        thirteen such cases, affecting characters in the following ranges:
1442    
1443          U+002b0 - U+002c1
1444          U+0060c - U+0060d
1445          U+0061e - U+00612
1446          U+0064b - U+0065e
1447          U+0074d - U+0076d
1448          U+01800 - U+01805
1449          U+01d00 - U+01d77
1450          U+01d9b - U+01dbf
1451          U+0200b - U+0200f
1452          U+030fc - U+030fe
1453          U+03260 - U+0327f
1454          U+0fb46 - U+0fbb1
1455          U+10450 - U+1049d
1456    
1457    12. The -o option (show only the matching part of a line) for pcregrep was not
1458        compatible with GNU grep in that, if there was more than one match in a
1459        line, it showed only the first of them. It now behaves in the same way as
1460        GNU grep.
1461    
1462    13. If the -o and -v options were combined for pcregrep, it printed a blank
1463        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1464        does the same. The return code can be used to tell if there were any
1465        non-matching lines.
1466    
1467    14. Added --file-offsets and --line-offsets to pcregrep.
1468    
1469    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1470        infinitely looping recursion. The bug was that positive lookaheads were not
1471        being skipped when checking for a possible empty match (negative lookaheads
1472        and both kinds of lookbehind were skipped).
1473    
1474    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1475        inclusion of <windows.h> to before rather than after the definition of
1476        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1477    
1478    17. Specifying a possessive quantifier with a specific limit for a Unicode
1479        character property caused pcre_compile() to compile bad code, which led at
1480        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1481        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1482        caused the error; without that there was no problem.
1483    
1484    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1485    
1486    19. Added --enable-pcretest-libreadline.
1487    
1488    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1489        RE::GlobalReplace(). As a result, the number of replacements returned was
1490        double what it should be. I removed one of the increments, but Craig sent a
1491        later patch that removed the other one (the right fix) and added unit tests
1492        that check the return values (which was not done before).
1493    
1494    21. Several CMake things:
1495    
1496        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1497            the names libpcre and libpcreposix, not just pcre and pcreposix.
1498    
1499        (2) The above change means that pcretest and pcregrep are now correctly
1500            linked with the newly-built libraries, not previously installed ones.
1501    
1502        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1503    
1504    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1505        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1506        UTF-8 newline character). The key issue is that the pattern starts .*;
1507        this means that the match must be either at the beginning, or after a
1508        newline. The bug was in the code for advancing after a failed match and
1509        checking that the new position followed a newline. It was not taking
1510        account of UTF-8 characters correctly.
1511    
1512    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1513        character classes. PCRE was not treating the sequence [:...:] as a
1514        character class unless the ... were all letters. Perl, however, seems to
1515        allow any characters between [: and :], though of course it rejects as
1516        unknown any "names" that contain non-letters, because all the known class
1517        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1518        for example, whereas PCRE did not - it did not recognize a POSIX character
1519        class. This seemed a bit dangerous, so the code has been changed to be
1520        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1521        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1522        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1523        Perl does, and where it didn't before.
1524    
1525    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1526        Windows environments %n is disabled by default.
1527    
1528    
1529    Version 7.4 21-Sep-07
1530    ---------------------
1531    
1532    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1533        means that a class such as [\s] counted as "explicit reference to CR or
1534        LF". That isn't really right - the whole point of the change was to try to
1535        help when there was an actual mention of one of the two characters. So now
1536        the change happens only if \r or \n (or a literal CR or LF) character is
1537        encountered.
1538    
1539    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1540        of both had grown to the point where there were only 3 bits left.
1541        Fortunately, there was spare space in the data structure, and so I have
1542        moved the internal flags into a new 16-bit field to free up more option
1543        bits.
1544    
1545    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1546        but did not set the internal JCHANGED flag - either of these is enough to
1547        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1548        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1549        start sets both bits.
1550    
1551    4.  Added options (at build time, compile time, exec time) to change \R from
1552        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1553    
1554    5.  doc/pcresyntax.html was missing from the distribution.
1555    
1556    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1557        compatibility, even though it is no longer used.
1558    
1559    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1560        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1561        windows.h file is present (where different names are used). [This was
1562        reversed later after testing - see 16 below.]
1563    
1564    8.  Changed all #include <config.h> to #include "config.h". There were also
1565        some further <pcre.h> cases that I changed to "pcre.h".
1566    
1567    9.  When pcregrep was used with the --colour option, it missed the line ending
1568        sequence off the lines that it output.
1569    
1570    10. It was pointed out to me that arrays of string pointers cause lots of
1571        relocations when a shared library is dynamically loaded. A technique of
1572        using a single long string with a table of offsets can drastically reduce
1573        these. I have refactored PCRE in four places to do this. The result is
1574        dramatic:
1575    
1576          Originally:                          290
1577          After changing UCP table:            187
1578          After changing error message table:   43
1579          After changing table of "verbs"       36
1580          After changing table of Posix names   22
1581    
1582        Thanks to the folks working on Gregex for glib for this insight.
1583    
1584    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1585        unicode-properties was also set.
1586    
1587    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1588    
1589    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1590        checked only for CRLF.
1591    
1592    14. Added casts to pcretest.c to avoid compiler warnings.
1593    
1594    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1595    
1596    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1597        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1598        entirely. This removes changes made in 7 above.
1599    
1600    17. The CMake files have been updated, and there is now more information about
1601        building with CMake in the NON-UNIX-USE document.
1602    
1603    
1604    Version 7.3 28-Aug-07
1605    ---------------------
1606    
1607     1. In the rejigging of the build system that eventually resulted in 7.1, the
1608        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1609        brackets there is not right, since it causes compilers to look for an
1610        installed pcre.h, not the version that is in the source that is being
1611        compiled (which of course may be different). I have changed it back to:
1612    
1613          #include "pcre.h"
1614    
1615        I have a vague recollection that the change was concerned with compiling in
1616        different directories, but in the new build system, that is taken care of
1617        by the VPATH setting the Makefile.
1618    
1619     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1620        when the subject happened to end in the byte 0x85 (e.g. if the last
1621        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1622        characters but of course it shouldn't be taken as a newline when it is part
1623        of another character. The bug was that, for an unlimited repeat of . in
1624        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1625        characters when looking for a newline.
1626    
1627     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1628    
1629     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1630        in debug output.
1631    
1632     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1633        long printing in the pcrecpp unittest when running under MinGW.
1634    
1635     6. ESC_K was left out of the EBCDIC table.
1636    
1637     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1638        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1639        limit also applies to "virtual nesting" when a pattern is recursive, and in
1640        this case 1000 isn't so big. I have been able to remove this limit at the
1641        expense of backing off one optimization in certain circumstances. Normally,
1642        when pcre_exec() would call its internal match() function recursively and
1643        immediately return the result unconditionally, it uses a "tail recursion"
1644        feature to save stack. However, when a subpattern that can match an empty
1645        string has an unlimited repetition quantifier, it no longer makes this
1646        optimization. That gives it a stack frame in which to save the data for
1647        checking that an empty string has been matched. Previously this was taken
1648        from the 1000-entry workspace that had been reserved. So now there is no
1649        explicit limit, but more stack is used.
1650    
1651     8. Applied Daniel's patches to solve problems with the import/export magic
1652        syntax that is required for Windows, and which was going wrong for the
1653        pcreposix and pcrecpp parts of the library. These were overlooked when this
1654        problem was solved for the main library.
1655    
1656     9. There were some crude static tests to avoid integer overflow when computing
1657        the size of patterns that contain repeated groups with explicit upper
1658        limits. As the maximum quantifier is 65535, the maximum group length was
1659        set at 30,000 so that the product of these two numbers did not overflow a
1660        32-bit integer. However, it turns out that people want to use groups that
1661        are longer than 30,000 bytes (though not repeat them that many times).
1662        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1663        made it possible to implement the integer overflow checks in a much more
1664        dynamic way, which I have now done. The artificial limitation on group
1665        length has been removed - we now have only the limit on the total length of
1666        the compiled pattern, which depends on the LINK_SIZE setting.
1667    
1668    10. Fixed a bug in the documentation for get/copy named substring when
1669        duplicate names are permitted. If none of the named substrings are set, the
1670        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1671        empty string.
1672    
1673    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1674        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1675        because the ] is interpreted as the first data character and the
1676        terminating ] is not found. PCRE has been made compatible with Perl in this
1677        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1678        cause memory overwriting.
1679    
1680    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1681        string has been matched (to stop an infinite loop). It was not recognizing
1682        a conditional subpattern that could match an empty string if that
1683        subpattern was within another subpattern. For example, it looped when
1684        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1685        condition was not nested. This bug has been fixed.
1686    
1687    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1688        past the start of the subject in the presence of bytes with the top bit
1689        set, for example "\x8aBCD".
1690    
1691    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1692        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1693    
1694    14. Optimized (?!) to (*FAIL).
1695    
1696    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1697        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1698        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1699        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1700        does: it's just the validity check that is more restrictive.
1701    
1702    16. Inserted checks for integer overflows during escape sequence (backslash)
1703        processing, and also fixed erroneous offset values for syntax errors during
1704        backslash processing.
1705    
1706    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1707        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1708    
1709    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1710        caused an overrun.
1711    
1712    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1713        something other than just ASCII characters) inside a group that had an
1714        unlimited repeat caused a loop at compile time (while checking to see
1715        whether the group could match an empty string).
1716    
1717    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1718        [\P{Any}] did so. (Error in the code for printing property names.)
1719    
1720    21. An orphan \E inside a character class could cause a crash.
1721    
1722    22. A repeated capturing bracket such as (A)? could cause a wild memory
1723        reference during compilation.
1724    
1725    23. There are several functions in pcre_compile() that scan along a compiled
1726        expression for various reasons (e.g. to see if it's fixed length for look
1727        behind). There were bugs in these functions when a repeated \p or \P was
1728        present in the pattern. These operators have additional parameters compared
1729        with \d, etc, and these were not being taken into account when moving along
1730        the compiled data. Specifically:
1731    
1732        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1733            length.
1734    
1735        (b) An item such as \pL+ within a repeated group could cause crashes or
1736            loops.
1737    
1738        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1739            "reference to non-existent subpattern" error.
1740    
1741        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1742    
1743    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1744        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1745    
1746    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1747    
1748    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1749        character were causing crashes (broken optimization).
1750    
1751    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1752        \p or \P) caused a compile-time loop.
1753    
1754    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1755        break. For example, the unstudied pattern [\r\n]A does not match the string
1756        "\r\nA" because change 7.0/46 below moves the current point on by two
1757        characters after failing to match at the start. However, the pattern \nA
1758        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1759        the same is true. There doesn't seem any very clean way out of this, but
1760        what I have chosen to do makes the common cases work: PCRE now takes note
1761        of whether there can be an explicit match for \r or \n anywhere in the
1762        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1763        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1764        pattern has explicit CR or LF references.
1765    
1766    29. Added (*CR) etc for changing newline setting at start of pattern.
1767    
1768    
1769    Version 7.2 19-Jun-07
1770    ---------------------
1771    
1772     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1773        which is apparently normally available under Windows.
1774    
1775     2. Re-jig the pcregrep tests with different newline settings in an attempt
1776        to make them independent of the local environment's newline setting.
1777    
1778     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1779    
1780     4. Some of the "internals" tests were previously cut out when the link size
1781        was not 2, because the output contained actual offsets. The recent new
1782        "Z" feature of pcretest means that these can be cut out, making the tests
1783        usable with all link sizes.
1784    
1785     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1786        stack recursion. This gives a massive performance boost under BSD, but just
1787        a small improvement under Linux. However, it saves one field in the frame
1788        in all cases.
1789    
1790     6. Added more features from the forthcoming Perl 5.10:
1791    
1792        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1793            recursion call. It refers to the nth most recently opened parentheses.
1794    
1795        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1796            to be opened parentheses.
1797    
1798        (c) Conditions that refer to capturing parentheses can be specified
1799            relatively, for example, (?(-2)... or (?(+3)...
1800    
1801        (d) \K resets the start of the current match so that everything before
1802            is not part of it.
1803    
1804        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1805    
1806        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1807            reference syntax.
1808    
1809        (g) (?| introduces a group in which the numbering of parentheses in each
1810            alternative starts with the same number.
1811    
1812        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1813    
1814     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1815        PCRE_INFO_JCHANGED.
1816    
1817     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1818        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1819        for detecting groups that can match an empty string.
1820    
1821     9. A pattern with a very large number of alternatives (more than several
1822        hundred) was running out of internal workspace during the pre-compile
1823        phase, where pcre_compile() figures out how much memory will be needed. A
1824        bit of new cunning has reduced the workspace needed for groups with
1825        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1826        workspace instead of running out of the 4096 that are available.
1827    
1828    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1829    
1830    11. Applied patch from Google to remove an optimization that didn't quite work.
1831        The report of the bug said:
1832    
1833          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1834          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1835          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1836    
1837    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1838        it matched the wrong number of bytes.
1839    
1840    
1841    Version 7.1 24-Apr-07
1842    ---------------------
1843    
1844     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1845        that is more "standard", making use of automake and other Autotools. There
1846        is some re-arrangement of the files and adjustment of comments consequent
1847        on this.
1848    
1849     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1850        for recursive directory scanning broke on some systems because the files
1851        are not scanned in any specific order and on different systems the order
1852        was different. A call to "sort" has been inserted into RunGrepTest for the
1853        approprate test as a short-term fix. In the longer term there may be an
1854        alternative.
1855    
1856     3. I had an email from Eric Raymond about problems translating some of PCRE's
1857        man pages to HTML (despite the fact that I distribute HTML pages, some
1858        people do their own conversions for various reasons). The problems
1859        concerned the use of low-level troff macros .br and .in. I have therefore
1860        removed all such uses from the man pages (some were redundant, some could
1861        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1862        HTML has been updated to handle .nf/.fi and to complain if it encounters
1863        .br or .in.
1864    
1865     4. Updated comments in configure.ac that get placed in config.h.in and also
1866        arranged for config.h to be included in the distribution, with the name
1867        config.h.generic, for the benefit of those who have to compile without
1868        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1869    
1870     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1871        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1872        makevp.bat for the current PCRE, using the additional files
1873        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1874    
1875     6. A Windows user reported a minor discrepancy with test 2, which turned out
1876        to be caused by a trailing space on an input line that had got lost in his
1877        copy. The trailing space was an accident, so I've just removed it.
1878    
1879     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1880        that is needed.
1881    
1882     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1883        as "const" (a) because they are and (b) because it helps the PHP
1884        maintainers who have recently made a script to detect big data structures
1885        in the php code that should be moved to the .rodata section. I remembered
1886        to update Builducptable as well, so it won't revert if ucptable.h is ever
1887        re-created.
1888    
1889     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1890        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1891        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1892        support is not required. This saves 1.5-2K of code, which is important in
1893        some applications.
1894    
1895        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1896        so as not to refer to the tables, even though these functions will never be
1897        called when UTF-8 support is disabled. Otherwise there are problems with a
1898        shared library.
1899    
1900    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1901    
1902        (a) It was defining its arguments as char * instead of void *.
1903    
1904        (b) It was assuming that all moves were upwards in memory; this was true
1905            a long time ago when I wrote it, but is no longer the case.
1906    
1907        The emulated memove() is provided for those environments that have neither
1908        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1909        is clearly not the case, as these two bugs were recently reported.
1910    
1911    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1912        and Detrail to create the HTML documentation, the .txt form of the man
1913        pages, and it removes trailing spaces from listed files. It also creates
1914        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1915        case, it wraps all the #defines with #ifndefs. This script should be run
1916        before "make dist".
1917    
1918    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1919        with Unicode property support.
1920    
1921        (a) For a maximizing quantifier, if the two different cases of the
1922            character were of different lengths in their UTF-8 codings (there are
1923            some cases like this - I found 11), and the matching function had to
1924            back up over a mixture of the two cases, it incorrectly assumed they
1925            were both the same length.
1926    
1927        (b) When PCRE was configured to use the heap rather than the stack for
1928            recursion during matching, it was not correctly preserving the data for
1929            the other case of a UTF-8 character when checking ahead for a match
1930            while processing a minimizing repeat. If the check also involved
1931            matching a wide character, but failed, corruption could cause an
1932            erroneous result when trying to check for a repeat of the original
1933            character.
1934    
1935    13. Some tidying changes to the testing mechanism:
1936    
1937        (a) The RunTest script now detects the internal link size and whether there
1938            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1939            values substituted by "configure". (The RunGrepTest script already did
1940            this for UTF-8.) The configure.ac script no longer substitutes the
1941            relevant variables.
1942    
1943        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1944            with length and offset values. This means that the output is different
1945            for different internal link sizes. Test 2 is skipped for link sizes
1946            other than 2 because of this, bypassing the problem. Unfortunately,
1947            there was also a test in test 3 (the locale tests) that used /B and
1948            failed for link sizes other than 2. Rather than cut the whole test out,
1949            I have added a new /Z option to pcretest that replaces the length and
1950            offset values with spaces. This is now used to make test 3 independent
1951            of link size. (Test 2 will be tidied up later.)
1952    
1953    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1954        segmentation fault instead of returning the appropriate error message.
1955    
1956    15. In multiline mode when the newline sequence was set to "any", the pattern
1957        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1958        This doesn't seem right; it now treats the CRLF combination as the line
1959        ending, and so does not match in that case. It's only a pattern such as ^$
1960        that would hit this one: something like ^ABC$ would have failed after \r
1961        and then tried again after \r\n.
1962    
1963    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1964        in an attempt to make files that differ only in their line terminators
1965        compare equal. This works on Linux.
1966    
1967    17. Under certain error circumstances pcregrep might try to free random memory
1968        as it exited. This is now fixed, thanks to valgrind.
1969    
1970    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1971        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1972        was because its rules for how to advance for /g after matching an empty
1973        string at the end of a line did not allow for this case. They now check for
1974        it specially.
1975    
1976    20. pcretest is supposed to handle patterns and data of any length, by
1977        extending its buffers when necessary. It was getting this wrong when the
1978        buffer for a data line had to be extended.
1979    
1980    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1981        CRLF as a newline sequence.
1982    
1983    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1984        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1985        I have nevertheless tidied it up.
1986    
1987    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1988    
1989    24. Added a man page for pcre-config.
1990    
1991    
1992    Version 7.0 19-Dec-06
1993    ---------------------
1994    
1995     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1996        moving to gcc 4.1.1.
1997    
1998     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1999        sys/time.h, which is documented as needed for this function. It doesn't
2000        seem to matter on Linux, but it showed up on some releases of OS X.
2001    
2002     3. It seems that there are systems where bytes whose values are greater than
2003        127 match isprint() in the "C" locale. The "C" locale should be the
2004        default when a C program starts up. In most systems, only ASCII printing
2005        characters match isprint(). This difference caused the output from pcretest
2006        to vary, making some of the tests fail. I have changed pcretest so that:
2007    
2008        (a) When it is outputting text in the compiled version of a pattern, bytes
2009            other than 32-126 are always shown as hex escapes.
2010    
2011        (b) When it is outputting text that is a matched part of a subject string,
2012            it does the same, unless a different locale has been set for the match
2013            (using the /L modifier). In this case, it uses isprint() to decide.
2014    
2015     4. Fixed a major bug that caused incorrect computation of the amount of memory
2016        required for a compiled pattern when options that changed within the
2017        pattern affected the logic of the preliminary scan that determines the
2018        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2019        that the computed length was too small. The symptoms of this bug were
2020        either the PCRE error "internal error: code overflow" from pcre_compile(),
2021        or a glibc crash with a message such as "pcretest: free(): invalid next
2022        size (fast)". Examples of patterns that provoked this bug (shown in
2023        pcretest format) are:
2024    
2025          /(?-x: )/x
2026          /(?x)(?-x: \s*#\s*)/
2027          /((?i)[\x{c0}])/8
2028          /(?i:[\x{c0}])/8
2029    
2030        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2031        is now done differently.
2032    
2033     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2034        wrapper classes; (b) implement a new function in the C++ scanner that is
2035        more efficient than the old way of doing things because it avoids levels of
2036        recursion in the regex matching; (c) add a paragraph to the documentation
2037        for the FullMatch() function.
2038    
2039     6. The escape sequence \n was being treated as whatever was defined as
2040        "newline". Not only was this contrary to the documentation, which states
2041        that \n is character 10 (hex 0A), but it also went horribly wrong when
2042        "newline" was defined as CRLF. This has been fixed.
2043    
2044     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2045        was being set to -1 for the "end of line" case (supposedly a value that no
2046        character can have). Though this value is never used (the check for end of
2047        line is "zero bytes in current character"), it caused compiler complaints.
2048        I've changed it to 0xffffffff.
2049    
2050     8. In pcre_version.c, the version string was being built by a sequence of
2051        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2052        string (as it is for production releases) called a macro with an empty
2053        argument. The C standard says the result of this is undefined. The gcc
2054        compiler treats it as an empty string (which was what was wanted) but it is
2055        reported that Visual C gives an error. The source has been hacked around to
2056        avoid this problem.
2057    
2058     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2059        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2060        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2061        of them did).
2062    
2063    10. Originally, pcretest opened its input and output without "b"; then I was
2064        told that "b" was needed in some environments, so it was added for release
2065        5.0 to both the input and output. (It makes no difference on Unix-like
2066        systems.) Later I was told that it is wrong for the input on Windows. I've
2067        now abstracted the modes into two macros, to make it easier to fiddle with
2068        them, and removed "b" from the input mode under Windows.
2069    
2070    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2071    
2072    12. Added -help and --help to pcretest as an official way of being reminded
2073        of the options.
2074    
2075    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2076        and pcrecpp.cc because they annoy compilers at high warning levels.
2077    
2078    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2079    
2080    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2081        scripts are not C programs :-) and which was not noticed because it works
2082        on Linux.
2083    
2084    16. pcretest is supposed to handle any length of pattern and data line (as one
2085        line or as a continued sequence of lines) by extending its input buffer if
2086        necessary. This feature was broken for very long pattern lines, leading to
2087        a string of junk being passed to pcre_compile() if the pattern was longer
2088        than about 50K.
2089    
2090    17. I have done a major re-factoring of the way pcre_compile() computes the
2091        amount of memory needed for a compiled pattern. Previously, there was code
2092        that made a preliminary scan of the pattern in order to do this. That was
2093        OK when PCRE was new, but as the facilities have expanded, it has become
2094        harder and harder to keep it in step with the real compile phase, and there
2095        have been a number of bugs (see for example, 4 above). I have now found a
2096        cunning way of running the real compile function in a "fake" mode that
2097        enables it to compute how much memory it would need, while actually only
2098        ever using a few hundred bytes of working memory and without too many
2099        tests of the mode. This should make future maintenance and development
2100        easier. A side effect of this work is that the limit of 200 on the nesting
2101        depth of parentheses has been removed (though this was never a serious
2102        limitation, I suspect). However, there is a downside: pcre_compile() now
2103        runs more slowly than before (30% or more, depending on the pattern). I
2104        hope this isn't a big issue. There is no effect on runtime performance.
2105    
2106    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2107        newline (only possible for the last line of a file) and it was a
2108        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2109    
2110    19. Added additional timing features to pcretest. (1) The -tm option now times
2111        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2112        separate command line item, by a number that specifies the number of
2113        repeats to use when timing. The default is 50000; this gives better
2114        precision, but takes uncomfortably long for very large patterns.
2115    
2116    20. Extended pcre_study() to be more clever in cases where a branch of a
2117        subpattern has no definite first character. For example, (a*|b*)[cd] would
2118        previously give no result from pcre_study(). Now it recognizes that the
2119        first character must be a, b, c, or d.
2120    
2121    21. There was an incorrect error "recursive call could loop indefinitely" if
2122        a subpattern (or the entire pattern) that was being tested for matching an
2123        empty string contained only one non-empty item after a nested subpattern.
2124        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2125        incorrectly, because the \d was being skipped in the check.
2126    
2127    22. The pcretest program now has a new pattern option /B and a command line
2128        option -b, which is equivalent to adding /B to every pattern. This causes
2129        it to show the compiled bytecode, without the additional information that
2130        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2131        is the same as /B/I).
2132    
2133    23. A new optimization is now able automatically to treat some sequences such
2134        as a*b as a*+b. More specifically, if something simple (such as a character
2135        or a simple class like \d) has an unlimited quantifier, and is followed by
2136        something that cannot possibly match the quantified thing, the quantifier
2137        is automatically "possessified".
2138    
2139    24. A recursive reference to a subpattern whose number was greater than 39
2140        went wrong under certain circumstances in UTF-8 mode. This bug could also
2141        have affected the operation of pcre_study().
2142    
2143    25. Realized that a little bit of performance could be had by replacing
2144        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2145    
2146    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2147    
2148    27. Possessive quantifiers such as a++ were previously implemented by turning
2149        them into atomic groups such as ($>a+). Now they have their own opcodes,
2150        which improves performance. This includes the automatically created ones
2151        from 23 above.
2152    
2153    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2154        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2155        the first matched character to be a colon. This applied both to named and
2156        numbered groups.
2157    
2158    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2159    
2160    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2161        building PCRE on Windows easier, so I have included it in the distribution.
2162    
2163    31. There is now a check in pcretest against a ridiculously large number being
2164        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2165        loop, the loop is abandoned.
2166    
2167    32. Forward references to subpatterns in conditions such as (?(2)...) where
2168        subpattern 2 is defined later cause pcre_compile() to search forwards in
2169        the pattern for the relevant set of parentheses. This search went wrong
2170        when there were unescaped parentheses in a character class, parentheses
2171        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2172    
2173    33. "Subroutine" calls and backreferences were previously restricted to
2174        referencing subpatterns earlier in the regex. This restriction has now
2175        been removed.
2176    
2177    34. Added a number of extra features that are going to be in Perl 5.10. On the
2178        whole, these are just syntactic alternatives for features that PCRE had
2179        previously implemented using the Python syntax or my own invention. The
2180        other formats are all retained for compatibility.
2181    
2182        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2183            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2184            also .NET compatible.
2185    
2186        (b) A recursion or subroutine call to a named group can now be defined as
2187            (?&name) as well as (?P>name).
2188    
2189        (c) A backreference to a named group can now be defined as \k<name> or
2190            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2191            5.10, are also .NET compatible.
2192    
2193        (d) A conditional reference to a named group can now use the syntax
2194            (?(<name>) or (?('name') as well as (?(name).
2195    
2196        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2197            groups (named and numbered) that are never evaluated inline, but can be
2198            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2199            is always false. There may be only one alternative in such a group.
2200    
2201        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2202            as the simple (?(R). The condition is true only if the most recent
2203            recursion is that of the given number or name. It does not search out
2204            through the entire recursion stack.
2205    
2206        (g) The escape \gN or \g{N} has been added, where N is a positive or
2207            negative number, specifying an absolute or relative reference.
2208    
2209    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2210        some "unreachable code" warnings.
2211    
2212    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2213        things, this adds five new scripts.
2214    
2215    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2216        There were also incompatibilities regarding the handling of \Q..\E inside
2217        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2218        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2219    
2220    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2221        matches an empty string, and forcibly breaks the loop. There were bugs in
2222        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2223        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2224        separate and independent bugs (that affected different cases) have been
2225        fixed.
2226    
2227    39. Refactored the code to abolish the use of different opcodes for small
2228        capturing bracket numbers. This is a tidy that I avoided doing when I
2229        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2230        The new approach is not only tidier, it makes it possible to reduce the
2231        memory needed to fix the previous bug (38).
2232    
2233    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2234        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2235        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2236        mode.
2237    
2238    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2239        report.
2240    
2241    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2242        copy construction and assignment in the C++ wrapper.
2243    
2244    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2245        couple of bugs that could have given wrong results in the "--newline=crlf"
2246        case.
2247    
2248    44. Added a number of casts and did some reorganization of signed/unsigned int
2249        variables following suggestions from Dair Grant. Also renamed the variable
2250        "this" as "item" because it is a C++ keyword.
2251    
2252    45. Arranged for dftables to add
2253    
2254          #include "pcre_internal.h"
2255    
2256        to pcre_chartables.c because without it, gcc 4.x may remove the array
2257        definition from the final binary if PCRE is built into a static library and
2258        dead code stripping is activated.
2259    
2260    46. For an unanchored pattern, if a match attempt fails at the start of a
2261        newline sequence, and the newline setting is CRLF or ANY, and the next two
2262        characters are CRLF, advance by two characters instead of one.
2263    
2264    
2265    Version 6.7 04-Jul-06
2266    ---------------------
2267    
2268     1. In order to handle tests when input lines are enormously long, pcretest has
2269        been re-factored so that it automatically extends its buffers when
2270        necessary. The code is crude, but this _is_ just a test program. The
2271        default size has been increased from 32K to 50K.
2272    
2273     2. The code in pcre_study() was using the value of the re argument before
2274        testing it for NULL. (Of course, in any sensible call of the function, it
2275        won't be NULL.)
2276    
2277     3. The memmove() emulation function in pcre_internal.h, which is used on
2278        systems that lack both memmove() and bcopy() - that is, hardly ever -
2279        was missing a "static" storage class specifier.
2280    
2281     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2282        containing an extended class (one that cannot be represented by a bitmap
2283        because it contains high-valued characters or Unicode property items, e.g.
2284        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2285        pattern, but PCRE should not loop if you do not (it no longer does).
2286        [Detail: two cases were found: (a) a repeated subpattern containing an
2287        extended class; (b) a recursive reference to a subpattern that followed a
2288        previous extended class. It wasn't skipping over the extended class
2289        correctly when UTF-8 mode was not set.]
2290    
2291     5. A negated single-character class was not being recognized as fixed-length
2292        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2293        compile error "lookbehind assertion is not fixed length".
2294    
2295     6. The RunPerlTest auxiliary script was showing an unexpected difference
2296        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2297        write a Perl script that can interpret lines of an input file either as
2298        byte characters or as UTF-8, which is what "perltest" was being required to
2299        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2300        can't do is switch easily at run time between having the "use utf8;" pragma
2301        or not. In the end, I fudged it by using the RunPerlTest script to insert
2302        "use utf8;" explicitly for the UTF-8 tests.
2303    
2304     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2305        the end of the subject string, contrary to the documentation and to what
2306        Perl does. This was true of both matching functions. Now it matches only at
2307        the start of the subject and immediately after *internal* newlines.
2308    
2309     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2310        a pointer to an int instead of a pointer to an unsigned long int. This
2311        caused problems on 64-bit systems.
2312    
2313     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2314        instance of the 'standard' template library not being so standard".
2315    
2316    10. There was no check on the number of named subpatterns nor the maximum
2317        length of a subpattern name. The product of these values is used to compute
2318        the size of the memory block for a compiled pattern. By supplying a very
2319        long subpattern name and a large number of named subpatterns, the size
2320        computation could be caused to overflow. This is now prevented by limiting
2321        the length of names to 32 characters, and the number of named subpatterns
2322        to 10,000.
2323    
2324    11. Subpatterns that are repeated with specific counts have to be replicated in
2325        the compiled pattern. The size of memory for this was computed from the
2326        length of the subpattern and the repeat count. The latter is limited to
2327        65535, but there was no limit on the former, meaning that integer overflow
2328        could in principle occur. The compiled length of a repeated subpattern is
2329        now limited to 30,000 bytes in order to prevent this.
2330    
2331    12. Added the optional facility to have named substrings with the same name.
2332    
2333    13. Added the ability to use a named substring as a condition, using the
2334        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2335        are numbers (not recommended). Forward references are permitted.
2336    
2337    14. Added forward references in named backreferences (if you see what I mean).
2338    
2339    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2340        pattern could run off the end of the subject. For example, the pattern
2341        "(?s)(.{1,5})"8 did this with the subject "ab".
2342    
2343    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2344        PCRE_CASELESS was set when matching characters that were quantified with ?
2345        or *.
2346    
2347    17. A character class other than a single negated character that had a minimum
2348        but no maximum quantifier - for example [ab]{6,} - was not handled
2349        correctly by pce_dfa_exec(). It would match only one character.
2350    
2351    18. A valid (though odd) pattern that looked like a POSIX character
2352        class but used an invalid character after [ (for example [[,abc,]]) caused
2353        pcre_compile() to give the error "Failed: internal error: code overflow" or
2354        in some cases to crash with a glibc free() error. This could even happen if
2355        the pattern terminated after [[ but there just happened to be a sequence of
2356        letters, a binary zero, and a closing ] in the memory that followed.
2357    
2358    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2359        over the years. Originally (before any Unicode support), just the bottom 8
2360        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2361        output from "man perlunicode" includes this:
2362    
2363          The regular expression compiler produces polymorphic opcodes.  That
2364          is, the pattern adapts to the data and automatically switches to
2365          the Unicode character scheme when presented with Unicode data--or
2366          instead uses a traditional byte scheme when presented with byte
2367          data.
2368    
2369        Sadly, a wide octal escape does not cause a switch, and in a string with
2370        no other multibyte characters, these octal escapes are treated as before.
2371        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2372        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2373        Unicode string.
2374    
2375        I have not perpetrated such confusion in PCRE. Up till now, it took just
2376        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2377        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2378        translate to the appropriate multibyte character.
2379    
2380    29. Applied some refactoring to reduce the number of warnings from Microsoft
2381        and Borland compilers. This has included removing the fudge introduced
2382        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2383        a warning about an unused variable.
2384    
2385    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2386        characters since release 4.0, because Perl (from release 5.004) does not.
2387        [Or at least, is documented not to: some releases seem to be in conflict
2388        with the documentation.] However, when a pattern was studied with
2389        pcre_study() and all its branches started with \s, PCRE still included VT
2390        as a possible starting character. Of course, this did no harm; it just
2391        caused an unnecessary match attempt.
2392    
2393    22. Removed a now-redundant internal flag bit that recorded the fact that case
2394        dependency changed within the pattern. This was once needed for "required
2395        byte" processing, but is no longer used. This recovers a now-scarce options
2396        bit. Also moved the least significant internal flag bit to the most-
2397        significant bit of the word, which was not previously used (hangover from
2398        the days when it was an int rather than a uint) to free up another bit for
2399        the future.
2400    
2401    23. Added support for CRLF line endings as well as CR and LF. As well as the
2402        default being selectable at build time, it can now be changed at runtime
2403        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2404        specify that it is scanning data with non-default line endings.
2405    
2406    24. Changed the definition of CXXLINK to make it agree with the definition of
2407        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2408    
2409    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2410        recursions. This makes a big different to stack usage for some patterns.
2411    
2412    26. If a subpattern containing a named recursion or subroutine reference such
2413        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2414        the space required for the compiled pattern went wrong and gave too small a
2415        value. Depending on the environment, this could lead to "Failed: internal
2416        error: code overflow at offset 49" or "glibc detected double free or
2417        corruption" errors.
2418    
2419    27. Applied patches from Google (a) to support the new newline modes and (b) to
2420        advance over multibyte UTF-8 characters in GlobalReplace.
2421    
2422    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2423        difference for some implementation of PCRE in some Windows version.
2424    
2425    29. Added some extra testing facilities to pcretest:
2426    
2427        \q<number>   in a data line sets the "match limit" value
2428        \Q<number>   in a data line sets the "match recursion limt" value
2429        -S <number>  sets the stack size, where <number> is in megabytes
2430    
2431        The -S option isn't available for Windows.
2432    
2433    
2434    Version 6.6 06-Feb-06
2435    ---------------------
2436    
2437     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2438        in pcreposix.h. I have copied the definition from pcre.h.
2439    
2440     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2441        because pcre.h is no longer a built file.
2442    
2443     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2444        not normally included in the compiled code.
2445    
2446    
2447    Version 6.5 01-Feb-06
2448    ---------------------
2449    
2450     1. When using the partial match feature with pcre_dfa_exec(), it was not
2451        anchoring the second and subsequent partial matches at the new starting
2452        point. This could lead to incorrect results. For example, with the pattern
2453        /1234/, partially matching against "123" and then "a4" gave a match.
2454    
2455     2. Changes to pcregrep:
2456    
2457        (a) All non-match returns from pcre_exec() were being treated as failures
2458            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2459            error message is output. Some extra information is given for the
2460            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2461            probably the only errors that are likely to be caused by users (by
2462            specifying a regex that has nested indefinite repeats, for instance).
2463            If there are more than 20 of these errors, pcregrep is abandoned.
2464    
2465        (b) A binary zero was treated as data while matching, but terminated the
2466            output line if it was written out. This has been fixed: binary zeroes
2467            are now no different to any other data bytes.
2468    
2469        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2470            used to set a locale for matching. The --locale=xxxx long option has
2471            been added (no short equivalent) to specify a locale explicitly on the
2472            pcregrep command, overriding the environment variables.
2473    
2474        (d) When -B was used with -n, some line numbers in the output were one less
2475            than they should have been.
2476    
2477        (e) Added the -o (--only-matching) option.
2478    
2479        (f) If -A or -C was used with -c (count only), some lines of context were
2480            accidentally printed for the final match.
2481    
2482        (g) Added the -H (--with-filename) option.
2483    
2484        (h) The combination of options -rh failed to suppress file names for files
2485            that were found from directory arguments.
2486    
2487        (i) Added the -D (--devices) and -d (--directories) options.
2488    
2489        (j) Added the -F (--fixed-strings) option.
2490    
2491        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2492    
2493        (l) Added the --colo(u)r option.
2494    
2495        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2496            is not present by default.
2497    
2498     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2499        items such as (?R) or (?1), when the recursion could match a number of
2500        alternatives. If it matched one of the alternatives, but subsequently,
2501        outside the recursion, there was a failure, the code tried to back up into
2502        the recursion. However, because of the way PCRE is implemented, this is not
2503        possible, and the result was an incorrect result from the match.
2504    
2505        In order to prevent this happening, the specification of recursion has
2506        been changed so that all such subpatterns are automatically treated as
2507        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2508    
2509     4. I had overlooked the fact that, in some locales, there are characters for
2510        which isalpha() is true but neither isupper() nor islower() are true. In
2511        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2512        and ordfeminine) are like this. This affected the treatment of \w and \W
2513        when they appeared in character classes, but not when they appeared outside
2514        a character class. The bit map for "word" characters is now created
2515        separately from the results of isalnum() instead of just taking it from the
2516        upper, lower, and digit maps. (Plus the underscore character, of course.)
2517    
2518     5. The above bug also affected the handling of POSIX character classes such as
2519        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2520        permanent tables. Instead, the bit maps for such a class were previously
2521        created as the appropriate unions of the upper, lower, and digit bitmaps.
2522        Now they are created by subtraction from the [[:word:]] class, which has
2523        its own bitmap.
2524    
2525     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2526        It is created by subtracting the vertical space characters (\x09, \x0a,
2527        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2528        subtraction was done in the overall bitmap for a character class, meaning
2529        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2530        be recognized. This bug has been fixed.
2531    
2532     7. Patches from the folks at Google:
2533    
2534          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2535          real life, but is still worth protecting against".
2536    
2537          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2538          regular expressions".
2539    
2540          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2541          have it.
2542    
2543          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2544          "configure" and the latter not, in order to fix a problem somebody had
2545          with compiling the Arg class on HP-UX.
2546    
2547          (e) Improve the error-handling of the C++ wrapper a little bit.
2548    
2549          (f) New tests for checking recursion limiting.
2550    
2551     8. The pcre_memmove() function, which is used only if the environment does not
2552        have a standard memmove() function (and is therefore rarely compiled),
2553        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2554        returning a result (though PCRE never actually uses the result).
2555    
2556     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2557        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2558        returned instead of calling malloc() with an overflowing number that would
2559        most likely cause subsequent chaos.
2560    
2561    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2562    
2563    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2564        with this option is matched, the nmatch and pmatch options of regexec() are
2565        ignored.
2566    
2567    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2568        provided in case anyone wants to the the POSIX interface with UTF-8
2569        strings.
2570    
2571    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2572        C++ linking (needed for some HP-UX environments).
2573    
2574    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2575        (unused parameter) and in the pcre_printint() function (omitted "default"
2576        switch label when the default is to do nothing).
2577    
2578    15. Added some code to make it possible, when PCRE is compiled as a C++
2579        library, to replace subject pointers for pcre_exec() with a smart pointer
2580        class, thus making it possible to process discontinuous strings.
2581    
2582    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2583        much the same function. They were added by different people who were trying
2584        to make PCRE easy to compile on non-Unix systems. It has been suggested
2585        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2586        for compiling on Windows systems. I have therefore replaced it with
2587        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2588        defaults to "extern" for C or "extern C" for C++, which works fine on
2589        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2590        SCOPE with something explicit in config.h. In addition:
2591    
2592        (a) pcreposix.h still had just "extern" instead of either of these macros;
2593            I have replaced it with PCRE_DATA_SCOPE.
2594    
2595        (b) Functions such as _pcre_xclass(), which are internal to the library,
2596            but external in the C sense, all had PCRE_EXPORT in their definitions.
2597            This is apparently wrong for the Windows case, so I have removed it.
2598            (It makes no difference on Unix-like systems.)
2599    
2600    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2601        of recursive calls to match(). This is different to MATCH_LIMIT because
2602        that limits the total number of calls to match(), not all of which increase
2603        the depth of recursion. Limiting the recursion depth limits the amount of
2604        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2605        when PCRE is compiled, and changed at run time. A patch from Google adds
2606        this functionality to the C++ interface.
2607    
2608    18. Changes to the handling of Unicode character properties:
2609    
2610        (a) Updated the table to Unicode 4.1.0.
2611    
2612        (b) Recognize characters that are not in the table as "Cn" (undefined).
2613    
2614        (c) I revised the way the table is implemented to a much improved format
2615            which includes recognition of ranges. It now supports the ranges that
2616            are defined in UnicodeData.txt, and it also amalgamates other
2617            characters into ranges. This has reduced the number of entries in the
2618            table from around 16,000 to around 3,000, thus reducing its size
2619            considerably. I realized I did not need to use a tree structure after
2620            all - a binary chop search is just as efficient. Having reduced the
2621            number of entries, I extended their size from 6 bytes to 8 bytes to
2622            allow for more data.
2623    
2624        (d) Added support for Unicode script names via properties such as \p{Han}.
2625    
2626    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2627        matching that character.
2628    
2629    20. When matching a repeated Unicode property with a minimum greater than zero,
2630        (for example \pL{2,}), PCRE could look past the end of the subject if it
2631        reached it while seeking the minimum number of characters. This could
2632        happen only if some of the characters were more than one byte long, because
2633        there is a check for at least the minimum number of bytes.
2634    
2635    21. Refactored the implementation of \p and \P so as to be more general, to
2636        allow for more different types of property in future. This has changed the
2637        compiled form incompatibly. Anybody with saved compiled patterns that use
2638        \p or \P will have to recompile them.
2639    
2640    22. Added "Any" and "L&" to the supported property types.
2641    
2642    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2643        but give a compile time error if the value is greater than 0xff.
2644    
2645    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2646        accidentally not being installed or uninstalled.
2647    
2648    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2649        made were to insert the current release number. This seemed silly, because
2650        it made things harder for people building PCRE on systems that don't run
2651        "configure". I have turned pcre.h into a distributed file, no longer built
2652        by "configure", with the version identification directly included. There is
2653        no longer a pcre.h.in file.
2654    
2655        However, this change necessitated a change to the pcre-config script as
2656        well. It is built from pcre-config.in, and one of the substitutions was the
2657        release number. I have updated configure.ac so that ./configure now finds
2658        the release number by grepping pcre.h.
2659    
2660    26. Added the ability to run the tests under valgrind.
2661    
2662    
2663    Version 6.4 05-Sep-05
2664    ---------------------
2665    
2666     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2667        "--" to be printed when multiple files were scanned, even when none of the
2668        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2669        consider it to be a bug, and have restored the previous behaviour.
2670    
2671     2. A couple of code tidies to get rid of compiler warnings.
2672    
2673     3. The pcretest program used to cheat by referring to symbols in the library
2674        whose names begin with _pcre_. These are internal symbols that are not
2675        really supposed to be visible externally, and in some environments it is
2676        possible to suppress them. The cheating is now confined to including
2677        certain files from the library's source, which is a bit cleaner.
2678    
2679     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2680        file's purpose clearer.
2681    
2682     5. Reorganized pcre_ucp_findchar().
2683    
2684    
2685    Version 6.3 15-Aug-05
2686    ---------------------
2687    
2688     1. The file libpcre.pc.in did not have general read permission in the tarball.
2689    
2690     2. There were some problems when building without C++ support:
2691    
2692        (a) If C++ support was not built, "make install" and "make test" still
2693            tried to test it.
2694    
2695        (b) There were problems when the value of CXX was explicitly set. Some
2696            changes have been made to try to fix these, and ...
2697    
2698        (c) --disable-cpp can now be used to explicitly disable C++ support.
2699    
2700        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2701            backslash in a target when C++ was disabled. This confuses some
2702            versions of "make", apparently. Using an intermediate variable solves
2703            this. (Same for CPP_LOBJ.)
2704    
2705     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2706        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2707        necessary on certain architectures.
2708    
2709     4. Added a setting of -export-symbols-regex to the link command to remove
2710        those symbols that are exported in the C sense, but actually are local
2711        within the library, and not documented. Their names all begin with
2712        "_pcre_". This is not a perfect job, because (a) we have to except some
2713        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2714        available (and never for static libraries). I have made a note to try to
2715        find a way round (a) in the future.
2716    
2717    
2718    Version 6.2 01-Aug-05
2719    ---------------------
2720    
2721     1. There was no test for integer overflow of quantifier values. A construction
2722        such as {1111111111111111} would give undefined results. What is worse, if
2723        a minimum quantifier for a parenthesized subpattern overflowed and became
2724        negative, the calculation of the memory size went wrong. This could have
2725        led to memory overwriting.
2726    
2727     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2728    
2729     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2730        operating environments where this matters.
2731    
2732     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2733        PCRE options from within the C++ wrapper.
2734    
2735     5. Named capturing subpatterns were not being correctly counted when a pattern
2736        was compiled. This caused two problems: (a) If there were more than 100
2737        such subpatterns, the calculation of the memory needed for the whole
2738        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2739        back references of the form \12, where the number was greater than 9, were
2740        not recognized as back references, even though there were sufficient
2741        previous subpatterns.
2742    
2743     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2744        versions of gcc, e.g. 2.95.4.
2745    
2746    
2747    Version 6.1 21-Jun-05
2748    ---------------------
2749    
2750     1. There was one reference to the variable "posix" in pcretest.c that was not
2751        surrounded by "#if !defined NOPOSIX".
2752    
2753     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2754        the cross-check on the old pcre_info() function, for the benefit of the
2755        cut-down version of PCRE that is currently imported into Exim.
2756    
2757     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2758        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2759        patterns that start (?i) but I don't think that matters. The use of (?i) is
2760        just an example; this all applies to the other options as well.
2761    
2762     4. Since libtool seems to echo the compile commands it is issuing, the output
2763        from "make" can be reduced a bit by putting "@" in front of each libtool
2764        compile command.
2765    
2766     5. Patch from the folks at Google for configure.in to be a bit more thorough
2767        in checking for a suitable C++ installation before trying to compile the
2768        C++ stuff. This should fix a reported problem when a compiler was present,
2769        but no suitable headers.
2770    
2771     6. The man pages all had just "PCRE" as their title. I have changed them to
2772        be the relevant file name. I have also arranged that these names are
2773        retained in the file doc/pcre.txt, which is a concatenation in text format
2774        of all the man pages except the little individual ones for each function.
2775    
2776     7. The NON-UNIX-USE file had not been updated for the different set of source
2777        files that come with release 6. I also added a few comments about the C++
2778        wrapper.
2779    
2780    
2781    Version 6.0 07-Jun-05
2782    ---------------------
2783    
2784     1. Some minor internal re-organization to help with my DFA experiments.
2785    
2786     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2787        didn't matter for the library itself when fully configured, but did matter
2788        when compiling without UCP support, or within Exim, where the ucp files are
2789        not imported.
2790    
2791     3. Refactoring of the library code to split up the various functions into
2792        different source modules. The addition of the new DFA matching code (see
2793        below) to a single monolithic source would have made it really too
2794        unwieldy, quite apart from causing all the code to be include in a
2795        statically linked application, when only some functions are used. This is
2796        relevant even without the DFA addition now that patterns can be compiled in
2797        one application and matched in another.
2798    
2799        The downside of splitting up is that there have to be some external
2800        functions and data tables that are used internally in different modules of
2801        the library but which are not part of the API. These have all had their
2802        names changed to start with "_pcre_" so that they are unlikely to clash
2803        with other external names.
2804    
2805     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2806        a different (DFA) algorithm. Although it is slower than the original
2807        function, it does have some advantages for certain types of matching
2808        problem.
2809    
2810     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2811        including restarting after a partial match.
2812    
2813     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2814        defined when compiling for Windows was sent to me. I have put it into the
2815        code, though I have no means of testing or verifying it.
2816    
2817     7. Added the pcre_refcount() auxiliary function.
2818    
2819     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2820        match before or at the first newline in the subject string. In pcretest,
2821        the /f option on a pattern can be used to set this.
2822    
2823     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2824        would behave wrongly. This has been present in PCRE since release 4.0.
2825    
2826    10. A number of changes to the pcregrep command:
2827    
2828        (a) Refactored how -x works; insert ^(...)$ instead of setting
2829            PCRE_ANCHORED and checking the length, in preparation for adding
2830            something similar for -w.
2831    
2832        (b) Added the -w (match as a word) option.
2833    
2834        (c) Refactored the way lines are read and buffered so as to have more
2835            than one at a time available.
2836    
2837        (d) Implemented a pcregrep test script.
2838    
2839        (e) Added the -M (multiline match) option. This allows patterns to match
2840            over several lines of the subject. The buffering ensures that at least
2841            8K, or the rest of the document (whichever is the shorter) is available
2842            for matching (and similarly the previous 8K for lookbehind assertions).
2843    
2844        (f) Changed the --help output so that it now says
2845    
2846              -w, --word-regex(p)
2847    
2848            instead of two lines, one with "regex" and the other with "regexp"
2849            because that confused at least one person since the short forms are the
2850            same. (This required a bit of code, as the output is generated
2851            automatically from a table. It wasn't just a text change.)
2852    
2853        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2854            option but starts with a hyphen. Could be a pattern or a path name
2855            starting with a hyphen, for instance.
2856    
2857        (h) "-" can be given as a file name to represent stdin.
2858    
2859        (i) When file names are being printed, "(standard input)" is used for
2860            the standard input, for compatibility with GNU grep. Previously
2861            "<stdin>" was used.
2862    
2863        (j) The option --label=xxx can be used to supply a name to be used for
2864            stdin when file names are being printed. There is no short form.
2865    
2866        (k) Re-factored the options decoding logic because we are going to add
2867            two more options that take data. Such options can now be given in four
2868            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2869    
2870        (l) Added the -A, -B, and -C options for requesting that lines of context
2871            around matches be printed.
2872    
2873        (m) Added the -L option to print the names of files that do not contain
2874            any matching lines, that is, the complement of -l.
2875    
2876        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2877            continue to scan other files.
2878    
2879        (o) The -s option was incorrectly implemented. For compatibility with other
2880            greps, it now suppresses the error message for a non-existent or non-
2881            accessible file (but not the return code). There is a new option called
2882            -q that suppresses the output of matching lines, which was what -s was
2883            previously doing.
2884    
2885        (p) Added --include and --exclude options to specify files for inclusion
2886            and exclusion when recursing.
2887    
2888    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2889        Hopefully, it now does.
2890    
2891    12. Missing cast in pcre_study().
2892    
2893    13. Added an "uninstall" target to the makefile.
2894    
2895    14. Replaced "extern" in the function prototypes in Makefile.in with
2896        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2897        world, but is set differently for Windows.
2898    
2899    15. Added a second compiling function called pcre_compile2(). The only
2900        difference is that it has an extra argument, which is a pointer to an
2901        integer error code. When there is a compile-time failure, this is set
2902        non-zero, in addition to the error test pointer being set to point to an
2903        error message. The new argument may be NULL if no error number is required
2904        (but then you may as well call pcre_compile(), which is now just a
2905        wrapper). This facility is provided because some applications need a
2906        numeric error indication, but it has also enabled me to tidy up the way
2907        compile-time errors are handled in the POSIX wrapper.
2908    
2909    16. Added VPATH=.libs to the makefile; this should help when building with one
2910        prefix path and installing with another. (Or so I'm told by someone who
2911        knows more about this stuff than I do.)
2912    
2913    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2914        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2915        match everything, including newlines. This is not POSIX-compatible, but
2916        somebody wanted the feature. From pcretest it can be activated by using
2917        both the P and the s flags.
2918    
2919    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2920    
2921    19. libpcre.pc was being incorrectly installed as executable.
2922    
2923    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2924        it now also looks for '\r' so that it will work unmodified on Windows.
2925    
2926    21. Added Google's contributed C++ wrapper to the distribution.
2927    
2928    22. Added some untidy missing memory free() calls in pcretest, to keep
2929        Electric Fence happy when testing.
2930    
2931    
2932    
2933    Version 5.0 13-Sep-04
2934    ---------------------
2935    
2936     1. Internal change: literal characters are no longer packed up into items
2937        containing multiple characters in a single byte-string. Each character
2938        is now matched using a separate opcode. However, there may be more than one
2939        byte in the character in UTF-8 mode.
2940    
2941     2. The pcre_callout_block structure has two new fields: pattern_position and
2942        next_item_length. These contain the offset in the pattern to the next match
2943        item, and its length, respectively.
2944    
2945     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2946        insertion of callouts before each pattern item. Added the /C option to
2947        pcretest to make use of this.
2948    
2949     4. On the advice of a Windows user, the lines
2950    
2951          #if defined(_WIN32) || defined(WIN32)
2952          _setmode( _fileno( stdout ), 0x8000 );
2953          #endif  /* defined(_WIN32) || defined(WIN32) */
2954    
2955        have been added to the source of pcretest. This apparently does useful
2956        magic in relation to line terminators.
2957    
2958     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2959        for the benefit of those environments where the "b" makes a difference.
2960    
2961     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2962        to know about it. I have put a hack into configure.in that adds in code
2963        to set GCC=yes if CC=icc. This seems to end up at a point in the
2964        generated configure script that is early enough to affect the setting of
2965        compiler options, which is what is needed, but I have no means of testing
2966        whether it really works. (The user who reported this had patched the
2967        generated configure script, which of course I cannot do.)
2968    
2969        LATER: After change 22 below (new libtool files), the configure script
2970        seems to know about icc (and also ecc). Therefore, I have commented out
2971        this hack in configure.in.
2972    
2973     7. Added support for pkg-config (2 patches were sent in).
2974    
2975     8. Negated POSIX character classes that used a combination of internal tables
2976        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2977        [[:^ascii]]. Typically, they would match almost any characters. The other
2978        POSIX classes were not broken in this way.
2979    
2980     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2981        to find the match, as PCRE was deluded into thinking that the match had to
2982        start at the start point or following a newline. The same bug applied to
2983        patterns with negative forward assertions or any backward assertions
2984        preceding ".*" at the start, unless the pattern required a fixed first
2985        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2986    
2987    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2988        starting at the last subject character, bytes beyond the end of the subject
2989        string were read.
2990    
2991    11. Renamed the variable "class" as "classbits" to make life easier for C++
2992        users. (Previously there was a macro definition, but it apparently wasn't
2993        enough.)
2994    
2995    12. Added the new field "tables" to the extra data so that tables can be passed
2996        in at exec time, or the internal tables can be re-selected. This allows
2997        a compiled regex to be saved and re-used at a later time by a different
2998        program that might have everything at different addresses.
2999    
3000    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3001        -R library as well as a -L library.
3002    
3003    14. The debugging options of pcretest (-d on the command line or D on a
3004        pattern) showed incorrect output for anything following an extended class
3005        that contained multibyte characters and which was followed by a quantifier.
3006    
3007    15. Added optional support for general category Unicode character properties
3008        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3009        support. It adds about 90K to the size of the library. The meanings of the
3010        inbuilt class escapes such as \d and \s have NOT been changed.
3011    
3012    16. Updated pcredemo.c to include calls to free() to release the memory for the
3013        compiled pattern.
3014    
3015    17. The generated file chartables.c was being created in the source directory
3016        instead of in the building directory. This caused the build to fail if the
3017        source directory was different from the building directory, and was
3018        read-only.
3019    
3020    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3021        file. No doubt somebody will tell me if they don't make sense... Also added
3022        Dan Mooney's comments about building on OpenVMS.
3023    
3024    19. Added support for partial matching via the PCRE_PARTIAL option for
3025        pcre_exec() and the \P data escape in pcretest.
3026    
3027    20. Extended pcretest with 3 new pattern features:
3028    
3029        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3030              write the compiled pattern to the file whose name is "rest-of-line".
3031              This is a straight binary dump of the data, with the saved pointer to
3032              the character tables forced to be NULL. The study data, if any, is
3033              written too. After writing, pcretest reads a new pattern.
3034    
3035        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3036              compiled pattern from the given file. There must not be any
3037              occurrences of "<" in the file name (pretty unlikely); if there are,
3038              pcretest will instead treat the initial "<" as a pattern delimiter.
3039              After reading in the pattern, pcretest goes on to read data lines as
3040              usual.
3041    
3042        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3043              and 16-bit fields in a compiled pattern, to simulate a pattern that
3044              was compiled on a host of opposite endianness.
3045    
3046    21. The pcre-exec() function can now cope with patterns that were compiled on
3047        hosts of opposite endianness, with this restriction:
3048    
3049          As for any compiled expression that is saved and used later, the tables
3050          pointer field cannot be preserved; the extra_data field in the arguments
3051          to pcre_exec() should be used to pass in a tables address if a value
3052          other than the default internal tables were used at compile time.
3053    
3054    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3055        now diagnosed as an error. Previously, most of the time, a negative number
3056        would have been treated as zero, but if in addition "ovector" was passed as
3057        NULL, a crash could occur.
3058    
3059    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3060        new versions from the libtool 1.5 distribution (the last one is a copy of
3061        a file called libtool.m4). This seems to have fixed the need to patch
3062        "configure" to support Darwin 1.3 (which I used to do). However, I still
3063        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3064        workstation).
3065    
3066    24. Changed the PCRE licence to be the more standard "BSD" licence.
3067    
3068    
3069    Version 4.5 01-Dec-03
3070    ---------------------
3071    
3072     1. There has been some re-arrangement of the code for the match() function so
3073        that it can be compiled in a version that does not call itself recursively.
3074        Instead, it keeps those local variables that need separate instances for
3075        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3076        needs to "recurse". Keeping track of where control must go is done by means
3077        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3078        hide most of the details from the main code, and operates only if
3079        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3080        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3081        operating.
3082    
3083        To make it easier for callers to provide specially tailored get/free
3084        functions for this usage, two new functions, pcre_stack_malloc, and
3085        pcre_stack_free, are used. They are always called in strict stacking order,
3086        and the size of block requested is always the same.
3087    
3088        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3089        PCRE has been compiled to use the stack or the heap for recursion. The
3090        -C option of pcretest uses this to show which version is compiled.
3091    
3092        A new data escape \S, is added to pcretest; it causes the amounts of store
3093        obtained and freed by both kinds of malloc/free at match time to be added
3094        to the output.
3095    
3096     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3097        what's available on my current Linux desktop machine.
3098    
3099     3. When matching a UTF-8 string, the test for a valid string at the start has
3100        been extended. If start_offset is not zero, PCRE now checks that it points
3101        to a byte that is the start of a UTF-8 character. If not, it returns
3102        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3103        this is necessary because there may be backward assertions in the pattern.
3104        When matching the same subject several times, it may save resources to use
3105        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3106    
3107     4. The code for checking the validity of UTF-8 strings has been tightened so
3108        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3109        containing "overlong sequences".
3110    
3111     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3112        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3113        should have been "&", but it just so happened that all the cases this let
3114        through by mistake were picked up later in the function.
3115    
3116     6. I had used a variable called "isblank" - this is a C99 function, causing
3117        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3118    
3119     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3120        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3121        so the version is shown; (c) stop "make test" echoing "./RunTest".
3122    
3123     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3124    
3125     9. The prototype for memmove() for systems that don't have it was using
3126        size_t, but the inclusion of the header that defines size_t was later. I've
3127        moved the #includes for the C headers earlier to avoid this.
3128    
3129    10. Added some adjustments to the code to make it easier to compiler on certain
3130        special systems:
3131    
3132          (a) Some "const" qualifiers were missing.
3133          (b) Added the macro EXPORT before all exported functions; by default this
3134              is defined to be empty.
3135          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3136              that it reads its output file name as an argument instead of writing
3137              to the standard output and assuming this can be redirected.
3138    
3139    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3140        class containing characters with values greater than 255, PCRE compilation
3141        went into a loop.
3142    
3143    12. A recursive reference to a subpattern that was within another subpattern
3144        that had a minimum quantifier of zero caused PCRE to crash. For example,
3145        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3146        recursion. If the recursively-called subpattern itself had a zero repeat,
3147        that was OK.
3148    
3149    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3150        buffer into which it was copied (for escape processing) was still set at
3151        1024, so long lines caused crashes.
3152    
3153    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3154        "internal error: code overflow...". This applied to any character class
3155        that was followed by a possessive quantifier.
3156    
3157    15. Modified the Makefile to add libpcre.la as a prerequisite for
3158        libpcreposix.la because I was told this is needed for a parallel build to
3159        work.
3160    
3161    16. If a pattern that contained .* following optional items at the start was
3162        studied, the wrong optimizing data was generated, leading to matching
3163        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3164        matching string must start with a or b or c. The correct conclusion for
3165        this pattern is that a match can start with any character.
3166    
3167    
3168    Version 4.4 13-Aug-03
3169    ---------------------
3170    
3171     1. In UTF-8 mode, a character class containing characters with values between
3172        127 and 255 was not handled correctly if the compiled pattern was studied.
3173        In fixing this, I have also improved the studying algorithm for such
3174        classes (slightly).
3175    
3176     2. Three internal functions had redundant arguments passed to them. Removal
3177        might give a very teeny performance improvement.
3178    
3179     3. Documentation bug: the value of the capture_top field in a callout is *one
3180        more than* the number of the hightest numbered captured substring.
3181    
3182     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3183        in incorrectly linking with a previously installed version. They now link
3184        explicitly with libpcre.la.
3185    
3186     5. configure.in no longer needs to recognize Cygwin specially.
3187    
3188     6. A problem in pcre.in for Windows platforms is fixed.
3189    
3190     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3191        pcretest, it used to include the size of the study block as part of its
3192        output. Unfortunately, the structure contains a field that has a different
3193        size on different hardware architectures. This meant that the tests that
3194        showed this size failed. As the block is currently always of a fixed size,
3195        this information isn't actually particularly useful in pcretest output, so
3196        I have just removed it.
3197    
3198     8. Three pre-processor statements accidentally did not start in column 1.
3199        Sadly, there are *still* compilers around that complain, even though
3200        standard C has not required this for well over a decade. Sigh.
3201    
3202     9. In pcretest, the code for checking callouts passed small integers in the
3203        callout_data field, which is a void * field. However, some picky compilers
3204        complained about the casts involved for this on 64-bit systems. Now
3205        pcretest passes the address of the small integer instead, which should get
3206        rid of the warnings.
3207    
3208    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3209        both compile and run time, and gives an error if an invalid UTF-8 sequence
3210        is found. There is a option for disabling this check in cases where the
3211        string is known to be correct and/or the maximum performance is wanted.
3212    
3213    11. In response to a bug report, I changed one line in Makefile.in from
3214    
3215            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3216        to
3217            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3218    
3219        to look similar to other lines, but I have no way of telling whether this
3220        is the right thing to do, as I do not use Windows. No doubt I'll get told
3221        if it's wrong...
3222    
3223    
3224    Version 4.3 21-May-03
3225    ---------------------
3226    
3227    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3228       Makefile.
3229    
3230    2. Some refactoring to improve the quality of the code:
3231    
3232       (i)   The utf8_table... variables are now declared "const".
3233    
3234       (ii)  The code for \cx, which used the "case flipping" table to upper case
3235             lower case letters, now just substracts 32. This is ASCII-specific,
3236             but the whole concept of \cx is ASCII-specific, so it seems
3237             reasonable.
3238    
3239       (iii) PCRE was using its character types table to recognize decimal and
3240             hexadecimal digits in the pattern. This is silly, because it handles
3241             only 0-9, a-f, and A-F, but the character types table is locale-
3242             specific, which means strange things might happen. A private
3243             table is now used for this - though it costs 256 bytes, a table is
3244             much faster than multiple explicit tests. Of course, the standard
3245             character types table is still used for matching digits in subject
3246             strings against \d.
3247    
3248       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3249             ending in _t are). So I've renamed it as ESC_tee.
3250    
3251    3. The first argument for regexec() in the POSIX wrapper should have been
3252       defined as "const".
3253    
3254    4. Changed pcretest to use malloc() for its buffers so that they can be
3255       Electric Fenced for debugging.
3256    
3257    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3258       to read one or more bytes before the start of the subject string. Often this
3259       had no effect on PCRE's behaviour, but in some circumstances it could
3260       provoke a segmentation fault.
3261    
3262    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3263       to try to read one or more bytes before the start of the subject string.
3264    
3265    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3266       UTF-8 support could misbehave in various ways if the subject string
3267       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3268       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3269       back over UTF-8 characters.)
3270    
3271    
3272    Version 4.2 14-Apr-03
3273    ---------------------
3274    
3275    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3276    
3277    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3278         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3279         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3280         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3281         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3282           and BUILD_EXEEXT
3283         Note: automatic setting of the BUILD variables is not yet working
3284         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3285           compile-time but not at link-time
3286         [LINK]: use for linking executables only
3287         make different versions for Windows and non-Windows
3288         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3289           libraries
3290         [LINK_FOR_BUILD]: new variable
3291         [OBJEXT]: use throughout
3292         [EXEEXT]: use throughout
3293         <winshared>: new target
3294         <wininstall>: new target
3295         <dftables.o>: use native compiler
3296         <dftables>: use native linker
3297         <install>: handle Windows platform correctly
3298         <clean>: ditto
3299         <check>: ditto
3300         copy DLL to top builddir before testing
3301    
3302       As part of these changes, -no-undefined was removed again. This was reported
3303       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3304       in any case.
3305    
3306    3. Some tidies to get rid of compiler warnings:
3307    
3308       . In the match_data structure, match_limit was an unsigned long int, whereas
3309         match_call_count was an int. I've made them both unsigned long ints.
3310    
3311       . In pcretest the fact that a const uschar * doesn't automatically cast to
3312         a void * provoked a warning.
3313    
3314       . Turning on some more compiler warnings threw up some "shadow" variables
3315         and a few more missing casts.
3316    
3317    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3318       option, a class that contained a single character with a value between 128
3319       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3320    
3321    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3322       option, a class that contained several characters, but with at least one
3323       whose value was between 128 and 255 caused PCRE to crash.
3324    
3325    
3326    Version 4.1 12-Mar-03
3327    ---------------------
3328    
3329    1. Compiling with gcc -pedantic found a couple of places where casts were
3330    needed, and a string in dftables.c that was longer than standard compilers are
3331    required to support.
3332    
3333    2. Compiling with Sun's compiler found a few more places where the code could
3334    be tidied up in order to avoid warnings.
3335    
3336    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3337    first of these names is deprecated in the latest Autoconf in favour of the name
3338    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3339    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3340    analogy I have changed it to CFLAGS_FOR_BUILD.
3341    
3342    4. Added -no-undefined to the linking command in the Makefile, because this is
3343    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3344    linking step for the pcreposix library.
3345    
3346    5. PCRE was failing to diagnose the case of two named groups with the same
3347    name.
3348    
3349    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3350    literal character that is needed in the subject for a match, and scans along to
3351    ensure that it is present before embarking on the full matching process. This
3352    saves time in cases of nested unlimited repeats that are never going to match.
3353    Problem: the scan can take a lot of time if the subject is very long (e.g.
3354    megabytes), thus penalizing straightforward matches. It is now done only if the
3355    amount of subject to be scanned is less than 1000 bytes.
3356    
3357    7. A lesser problem with the same optimization is that it was recording the
3358    first character of an anchored pattern as "needed", thus provoking a search
3359    right along the subject, even when the first match of the pattern was going to
3360    fail. The "needed" character is now not set for anchored patterns, unless it
3361    follows something in the pattern that is of non-fixed length. Thus, it still
3362    fulfils its original purpose of finding quick non-matches in cases of nested
3363    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3364    
3365    
3366    Version 4.0 17-Feb-03
3367    ---------------------
3368    
3369    1. If a comment in an extended regex that started immediately after a meta-item
3370    extended to the end of string, PCRE compiled incorrect data. This could lead to
3371    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3372    
3373    2. Moved to autoconf 2.53 and libtool 1.4.2.
3374    
3375    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3376    the special perltest8 script is no longer needed - all the tests can be run
3377    from a single perltest script.
3378    
3379    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3380    by \s. It has now been removed in PCRE. This means it isn't recognized as
3381    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3382    class [:space:] *does* include VT, thereby creating a mess.
3383    
3384    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3385    space and tab.
3386    
3387    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3388    its new features into the main test script, reducing the number of scripts.
3389    
3390    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3391    were backward compatible, and made the (?i) apply to the whole pattern, as if
3392    /i were given. Now it behaves more logically, and applies the option setting
3393    only to what follows. PCRE has been changed to follow suit. However, if it
3394    finds options settings right at the start of the pattern, it extracts them into
3395    the global options, as before. Thus, they show up in the info data.
3396    
3397    8. Added support for the \Q...\E escape sequence. Characters in between are
3398    treated as literals. This is slightly different from Perl in that $ and @ are
3399    also handled as literals inside the quotes. In Perl, they will cause variable
3400    interpolation. Note the following examples:
3401    
3402        Pattern            PCRE matches      Perl matches
3403    
3404        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3405        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3406        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3407    
3408    For compatibility with Perl, \Q...\E sequences are recognized inside character
3409    classes as well as outside them.
3410    
3411    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3412    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3413    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3414    signed/unsigned warnings.
3415    
3416    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3417    option for pcretest, so I've replaced it by a simple function that does just
3418    that job.
3419    
3420    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3421    "pcregrep -".
3422    
3423    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3424    Java package. This provides some syntactic sugar for simple cases of what my
3425    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3426    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3427    item, you can use this simplified notation. Note that only makes sense with
3428    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3429    greediness, whatever the setting of the PCRE_UNGREEDY option.
3430    
3431    13. A change of greediness default within a pattern was not taking effect at
3432    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3433    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3434    was abstracted outside.
3435    
3436    14. PCRE now supports the \G assertion. It is true when the current matching
3437    position is at the start point of the match. This differs from \A when the
3438    starting offset is non-zero. Used with the /g option of pcretest (or similar
3439    code), it works in the same way as it does for Perl's /g option. If all
3440    alternatives of a regex begin with \G, the expression is anchored to the start
3441    match position, and the "anchored" flag is set in the compiled expression.
3442    
3443    15. Some bugs concerning the handling of certain option changes within patterns
3444    have been fixed. These applied to options other than (?ims). For example,
3445    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3446    been the other way round. Some of this was related to change 7 above.
3447    
3448    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3449    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3450    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3451    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3452    
3453    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3454    mode. Unlike ".", it always matches newline, whatever the setting of
3455    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3456    assertions. Perl allows it, but it doesn't (in general) work because it can't
3457    calculate the length of the lookbehind. At least, that's the case for Perl
3458    5.8.0 - I've been told they are going to document that it doesn't work in
3459    future.
3460    
3461    18. Added an error diagnosis for escapes that PCRE does not support: these are
3462    \L, \l, \N, \P, \p, \U, \u, and \X.
3463    
3464    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3465    reading past the end of the pattern in cases such as /[abcd/.
3466    
3467    20. PCRE was getting more memory than necessary for patterns with classes that
3468    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3469    
3470    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3471    compiling PCRE for use with Virtual Pascal.
3472    
3473    22. Small fix to the Makefile to make it work properly if the build is done
3474    outside the source tree.
3475    
3476    23. Added a new extension: a condition to go with recursion. If a conditional
3477    subpattern starts with (?(R) the "true" branch is used if recursion has
3478    happened, whereas the "false" branch is used only at the top level.
3479    
3480    24. When there was a very long string of literal characters (over 255 bytes
3481    without UTF support, over 250 bytes with UTF support), the computation of how
3482    much memory was required could be incorrect, leading to segfaults or other
3483    strange effects.
3484    
3485    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3486    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3487    there was a subsequent back reference to those brackets. This meant that, for
3488    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3489    possible to check for precisely this case. All we can do is abandon the
3490    optimization if .* occurs inside capturing brackets when there are any back
3491    references whatsoever. (See below for a better fix that came later.)
3492    
3493    26. The handling of the optimization for finding the first character of a
3494    non-anchored pattern, and for finding a character that is required later in the
3495    match were failing in some cases. This didn't break the matching; it just
3496    failed to optimize when it could. The way this is done has been re-implemented.
3497    
3498    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3499    
3500    28. Added a new feature that provides some of the functionality that Perl
3501    provides with (?{...}). The facility is termed a "callout". The way it is done
3502    in PCRE is for the caller to provide an optional function, by setting
3503    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3504    global variable. By default it is unset, which disables all calling out. To get
3505    the function called, the regex must include (?C) at appropriate points. This
3506    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3507    This provides a means of identifying different callout points. When PCRE
3508    reaches such a point in the regex, if pcre_callout has been set, the external
3509    function is called. It is provided with data in a structure called
3510    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3511    matching continues; if it returns a non-zero value, the match at the current
3512    point fails. However, backtracking will occur if possible. [This was changed
3513    later and other features added - see item 49 below.]
3514    
3515    29. pcretest is upgraded to test the callout functionality. It provides a
3516    callout function that displays information. By default, it shows the start of
3517    the match and the current position in the text. There are some new data escapes
3518    to vary what happens:
3519    
3520        \C+         in addition, show current contents of captured substrings
3521        \C-         do not supply a callout function
3522        \C!n        return 1 when callout number n is reached
3523        \C!n!m      return 1 when callout number n is reached for the mth time
3524    
3525    30. If pcregrep was called with the -l option and just a single file name, it
3526    output "<stdin>" if a match was found, instead of the file name.
3527    
3528    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3529    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3530    pcre_exec(). This saves a malloc/free per call. The default value of
3531    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3532    when configuring.
3533    
3534    32. The default maximum size of a compiled pattern is 64K. There have been a
3535    few cases of people hitting this limit. The code now uses macros to handle the
3536    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3537    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3538    configuring. Tests 2 and 5 work only with 2-byte links because they output
3539    debugging information about compiled patterns.
3540    
3541    33. Internal code re-arrangements:
3542    
3543    (a) Moved the debugging function for printing out a compiled regex into
3544        its own source file (printint.c) and used #include to pull it into
3545        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3546        separate copies.
3547    
3548    (b) Defined the list of op-code names for debugging as a macro in
3549        internal.h so that it is next to the definition of the opcodes.
3550    
3551    (c) Defined a table of op-code lengths for simpler skipping along compiled
3552        code. This is again a macro in internal.h so that it is next to the
3553        definition of the opcodes.
3554    
3555    34. Added support for recursive calls to individual subpatterns, along the
3556    lines of Robin Houston's patch (but implemented somewhat differently).
3557    
3558    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3559    allow it to read and process whole directories in Win32. This code was
3560    contributed by Lionel Fourquaux; it has not been tested by me.
3561    
3562    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3563    used to name a group. Names consist of alphanumerics and underscores, and must
3564    be unique. Back references use the syntax (?P=name) and recursive calls use
3565    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3566    numbers. The function pcre_fullinfo() can be used after compilation to extract
3567    a name/number map. There are three relevant calls:
3568    
3569      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3570      PCRE_INFO_NAMECOUNT            yields the number of entries
3571      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3572    
3573    The map is a vector of fixed-size entries. The size of each entry depends on
3574    the length of the longest name used. The first two bytes of each entry are the
3575    group number, most significant byte first. There follows the corresponding
3576    name, zero terminated. The names are in alphabetical order.
3577    
3578    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3579    case instead of 255. Making it the same both with and without UTF-8 support
3580    means that the same test output works with both.
3581    
3582    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3583    calling malloc() with a zero argument.
3584    
3585    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3586    optimization. I've improved things by keeping a bitmap of backreferences with
3587    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3588    fact referenced, the optimization can be applied. It is unlikely that a
3589    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3590    the match to follow \n) will appear inside brackets with a number greater than
3591    31, but if it does, any back reference > 31 suppresses the optimization.
3592    
3593    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3594    of disabling numbered capturing parentheses. Any opening parenthesis that is
3595    not followed by ? behaves as if it were followed by ?: but named parentheses
3596    can still be used for capturing (and they will acquire numbers in the usual
3597    way).
3598    
3599    41. Redesigned the return codes from the match() function into yes/no/error so
3600    that errors can be passed back from deep inside the nested calls. A malloc
3601    failure while inside a recursive subpattern call now causes the
3602    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3603    
3604    42. It is now possible to set a limit on the number of times the match()
3605    function is called in a call to pcre_exec(). This facility makes it possible to
3606    limit the amount of recursion and backtracking, though not in a directly
3607    obvious way, because the match() function is used in a number of different
3608    circumstances. The count starts from zero for each position in the subject
3609    string (for non-anchored patterns). The default limit is, for compatibility, a
3610    large number, namely 10 000 000. You can change this in two ways:
3611    
3612    (a) When configuring PCRE before making, you can use --with-match-limit=n
3613        to set a default value for the compiled library.
3614    
3615    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3616        a different value is set. See 45 below.
3617    
3618    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3619    
3620    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3621    of things that can be changed at compile time. The first argument specifies
3622    what is wanted and the second points to where the information is to be placed.
3623    The current list of available information is:
3624    
3625      PCRE_CONFIG_UTF8
3626    
3627    The output is an integer that is set to one if UTF-8 support is available;
3628    otherwise it is set to zero.
3629    
3630      PCRE_CONFIG_NEWLINE
3631    
3632    The output is an integer that it set to the value of the code that is used for
3633    newline. It is either LF (10) or CR (13).
3634    
3635      PCRE_CONFIG_LINK_SIZE
3636    
3637    The output is an integer that contains the number of bytes used for internal
3638    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3639    
3640      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3641    
3642    The output is an integer that contains the threshold above which the POSIX
3643    interface uses malloc() for output vectors. See item 31 above.
3644    
3645      PCRE_CONFIG_MATCH_LIMIT
3646    
3647    The output is an unsigned integer that contains the default limit of the number
3648    of match() calls in a pcre_exec() execution. See 42 above.
3649    
3650    44. pcretest has been upgraded by the addition of the -C option. This causes it
3651    to extract all the available output from the new pcre_config() function, and to
3652    output it. The program then exits immediately.
3653    
3654    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3655    order to support additional features. One way would have been to define
3656    pcre_exec2() (for example) with extra arguments, but this would not have been
3657    extensible, and would also have required all calls to the original function to
3658    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3659    is used for passing in "extra" data from pcre_study().
3660    
3661    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3662    contains the following fields:
3663    
3664      flags         a bitmap indicating which of the following fields are set
3665      study_data    opaque data from pcre_study()
3666      match_limit   a way of specifying a limit on match() calls for a specific
3667                      call to pcre_exec()
3668      callout_data  data for callouts (see 49 below)
3669    
3670    The flag bits are also defined in pcre.h, and are
3671    
3672      PCRE_EXTRA_STUDY_DATA
3673      PCRE_EXTRA_MATCH_LIMIT
3674      PCRE_EXTRA_CALLOUT_DATA
3675    
3676    The pcre_study() function now returns one of these new pcre_extra blocks, with
3677    the actual study data pointed to by the study_data field, and the
3678    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3679    before. That is, this change is entirely upwards-compatible and requires no
3680    change to existing code.
3681    
3682    If you want to pass in additional data to pcre_exec(), you can either place it
3683    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3684    block.
3685    
3686    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3687    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3688    times with different match limits, until it finds the minimum value needed for
3689    pcre_exec() to complete. The value is then output. This can be instructive; for
3690    most simple matches the number is quite small, but for pathological cases it
3691    gets very large very quickly.
3692    
3693    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3694    returns the size of the data block pointed to by the study_data field in a
3695    pcre_extra block, that is, the value that was passed as the argument to
3696    pcre_malloc() when PCRE was getting memory in which to place the information
3697    created by pcre_study(). The fourth argument should point to a size_t variable.
3698    pcretest has been extended so that this information is shown after a successful
3699    pcre_study() call when information about the compiled regex is being displayed.
3700    
3701    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3702    because what follows is always an absolute path. (Later: it turns out that this
3703    is more than cosmetic for MinGW, because it doesn't like empty path
3704    components.)
3705    
3706    49. Some changes have been made to the callout feature (see 28 above):
3707    
3708    (i)  A callout function now has three choices for what it returns:
3709    
3710           0  =>  success, carry on matching
3711         > 0  =>  failure at this point, but backtrack if possible
3712         < 0  =>  serious error, return this value from pcre_exec()
3713    
3714         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3715         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3716         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3717         use by callout functions. It will never be used by PCRE itself.
3718    
3719    (ii) The pcre_extra structure (see 45 above) has a void * field called
3720         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3721         pcre_callout_block structure has a field of the same name. The contents of
3722         the field passed in the pcre_extra structure are passed to the callout
3723         function in the corresponding field in the callout block. This makes it
3724         easier to use the same callout-containing regex from multiple threads. For
3725         testing, the pcretest program has a new data escape
3726    
3727           \C*n        pass the number n (may be negative) as callout_data
3728    
3729         If the callout function in pcretest receives a non-zero value as
3730         callout_data, it returns that value.
3731    
3732    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3733    there were some redundant $(CFLAGS) in commands that are now specified as
3734    $(LINK), which already includes $(CFLAGS).
3735    
3736    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3737    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3738    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3739    one-byte characters throughout. Note that case-insensitive matching applies
3740    only to characters whose values are less than 256. PCRE doesn't support the
3741    notion of cases for higher-valued characters.
3742    
3743    (i)   A character class whose characters are all within 0-255 is handled as
3744          a bit map, and the map is inverted for negative classes. Previously, a
3745          character > 255 always failed to match such a class; however it should
3746          match if the class was a negative one (e.g. [^ab]). This has been fixed.
3747    
3748    (ii)  A negated character class with a single character < 255 is coded as
3749          "not this character" (OP_NOT). This wasn't working properly when the test
3750          character was multibyte, either singly or repeated.
3751    
3752    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3753          mode, for example: \x{100}{2,3}.
3754    
3755    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3756          singly or repeated) now correctly test multibyte characters. However,
3757          PCRE doesn't recognize any characters with values greater than 255 as
3758          digits, spaces, or word characters. Such characters always match \D, \S,
3759          and \W, and never match \d, \s, or \w.
3760    
3761    (v)   Classes may now contain characters and character ranges with values
3762          greater than 255. For example: [ab\x{100}-\x{400}].
3763    
3764    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
3765          PCRE in UTF-8 mode.
3766    
3767    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3768    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3769    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3770    value.)
3771    
3772    53. The single man page has become too large. I have therefore split it up into
3773    a number of separate man pages. These also give rise to individual HTML pages;
3774    these are now put in a separate directory, and there is an index.html page that
3775    lists them all. Some hyperlinking between the pages has been installed.
3776    
3777    54. Added convenience functions for handling named capturing parentheses.
3778    
3779    55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3780    aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3781    true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3782    are faulted.
3783    
3784    56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3785    calling configure. These values are used when compiling the dftables.c program
3786    which is run to generate the source of the default character tables. They
3787    default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3788    you will need to set these values.
3789    
3790    57. Updated the building process for Windows DLL, as provided by Fred Cox.
3791    
3792    
3793    Version 3.9 02-Jan-02
3794    ---------------------
3795    
3796    1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3797    
3798    2. If --disable-static was given, the building process failed when trying to
3799    build pcretest and pcregrep. (For some reason it was using libtool to compile
3800    them, which is not right, as they aren't part of the library.)
3801    
3802    
3803    Version 3.8 18-Dec-01
3804    ---------------------
3805    
3806    1. The experimental UTF-8 code was completely screwed up. It was packing the
3807    bytes in the wrong order. How dumb can you get?
3808    
3809    
3810    Version 3.7 29-Oct-01
3811    ---------------------
3812