/[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 882 by ph10, Sun Jan 15 18:45:27 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.30
5    ------------
6    
7    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
8        name is defined in ctype.h.
9    
10    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
11        only in quite long subpatterns.
12    
13    3.  Removed the function pcre_info(), which has been obsolete and deprecated
14        since it was replaced by pcre_fullinfo() in February 2000.
15    
16    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
17        match a (*MARK), and the match failed at the start of the subject, a
18        reference to memory before the start of the subject could occur. This bug
19        was introduced by fix 17 of release 8.21.
20    
21    5.  A reference to an unset group with zero minimum repetition was giving
22        totally wrong answers (in non-JavaScript-compatibility mode). For example,
23        /(another)?(\1?)test/ matched against "hello world test". This bug was
24        introduced in release 8.13.
25    
26    6.  Add support for 16-bit character strings (a large amount of work involving
27        many changes and refactorings).
28    
29    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
30        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
31        from a file.
32    
33    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
34        rounding is not applied in this particular case).
35    
36    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
37        if they appear, or are escaped, in patterns.
38    
39    10. Get rid of a number of -Wunused-but-set-variable warnings.
40    
41    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
42        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
43        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
44        also returns the mark "x". This bug applied to capturing parentheses,
45        non-capturing parentheses, and atomic parentheses.
46    
47    
48    Version 8.21 12-Dec-2011
49    ------------------------
50    
51    1.  Updating the JIT compiler.
52    
53    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
54        are added as well.
55    
56    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
57        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
58        calling _pcre_jit_exec. Some extra comments are added.
59    
60    4.  (*MARK) settings inside atomic groups that do not contain any capturing
61        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
62        was introduced by change 18 for 8.20.
63    
64    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
65        ECMA-262 standard.
66    
67    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
68        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
69        This bug was probably introduced by change 9 of 8.13.
70    
71    7.  While fixing 6 above, I noticed that a number of other items were being
72        incorrectly rejected as "not fixed length". This arose partly because newer
73        opcodes had not been added to the fixed-length checking code. I have (a)
74        corrected the bug and added tests for these items, and (b) arranged for an
75        error to occur if an unknown opcode is encountered while checking for fixed
76        length instead of just assuming "not fixed length". The items that were
77        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
78        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
79        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
80    
81    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
82        being incorrectly compiled and would have given unpredicatble results.
83    
84    9.  A possessively repeated subpattern with minimum repeat count greater than
85        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
86        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
87        the first (A) could occur when it should not.
88    
89    10. Add a cast and remove a redundant test from the code.
90    
91    11. JIT should use pcre_malloc/pcre_free for allocation.
92    
93    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
94        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
95        is anything other than /usr, -L is still shown).
96    
97    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
98    
99    14. Perl does not support \N without a following name in a [] class; PCRE now
100        also gives an error.
101    
102    15. If a forward reference was repeated with an upper limit of around 2000,
103        it caused the error "internal error: overran compiling workspace". The
104        maximum number of forward references (including repeats) was limited by the
105        internal workspace, and dependent on the LINK_SIZE. The code has been
106        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
107        the default depends on LINK_SIZE. There is a new upper limit (for safety)
108        of around 200,000 forward references. While doing this, I also speeded up
109        the filling in of repeated forward references.
110    
111    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
112        incorrectly expecting the subject to contain another "a" after the start.
113    
114    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
115        in the match, the SKIP should be ignored. This was not happening; instead
116        the SKIP was being treated as NOMATCH. For patterns such as
117        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
118        tested.
119    
120    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
121        now much more compatible with Perl, in particular in cases where the result
122        is a non-match for a non-anchored pattern. For example, if
123        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
124        "m", where previously it did not return a name. A side effect of this
125        change is that for partial matches, the last encountered mark name is
126        returned, as for non matches. A number of tests that were previously not
127        Perl-compatible have been moved into the Perl-compatible test files. The
128        refactoring has had the pleasing side effect of removing one argument from
129        the match() function, thus reducing its stack requirements.
130    
131    19. If the /S+ option was used in pcretest to study a pattern using JIT,
132        subsequent uses of /S (without +) incorrectly behaved like /S+.
133    
134    21. Retrieve executable code size support for the JIT compiler and fixing
135        some warnings.
136    
137    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
138        not work when the shorter character appeared right at the end of the
139        subject string.
140    
141    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
142        systems.
143    
144    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
145        output it when the /M option is used in pcretest.
146    
147    25. The CheckMan script was not being included in the distribution. Also, added
148        an explicit "perl" to run Perl scripts from the PrepareRelease script
149        because this is reportedly needed in Windows.
150    
151    26. If study data was being save in a file and studying had not found a set of
152        "starts with" bytes for the pattern, the data written to the file (though
153        never used) was taken from uninitialized memory and so caused valgrind to
154        complain.
155    
156    27. Updated RunTest.bat as provided by Sheri Pierce.
157    
158    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
159    
160    29. Computation of memory usage for the table of capturing group names was
161        giving an unnecessarily large value.
162    
163    
164    Version 8.20 21-Oct-2011
165    ------------------------
166    
167    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
168        a POSIX class. After further experiments with Perl, which convinced me that
169        Perl has bugs and confusions, a closing square bracket is no longer allowed
170        in a POSIX name. This bug also affected patterns with classes that started
171        with full stops.
172    
173    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
174        captured substring, but while checking the failing first alternative,
175        substring 1 is temporarily captured. If the output vector supplied to
176        pcre_exec() was not big enough for this capture, the yield of the function
177        was still zero ("insufficient space for captured substrings"). This cannot
178        be totally fixed without adding another stack variable, which seems a lot
179        of expense for a edge case. However, I have improved the situation in cases
180        such as /(a)(b)x|abc/ matched against "abc", where the return code
181        indicates that fewer than the maximum number of slots in the ovector have
182        been set.
183    
184    3.  Related to (2) above: when there are more back references in a pattern than
185        slots in the output vector, pcre_exec() uses temporary memory during
186        matching, and copies in the captures as far as possible afterwards. It was
187        using the entire output vector, but this conflicts with the specification
188        that only 2/3 is used for passing back captured substrings. Now it uses
189        only the first 2/3, for compatibility. This is, of course, another edge
190        case.
191    
192    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
193        main code base, and can be used by building with --enable-jit. When this is
194        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
195        runtime --no-jit option is given.
196    
197    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
198        ovector, the return from the function was zero, implying that there were
199        other matches that did not fit. The correct "exactly full" value is now
200        returned.
201    
202    6.  If a subpattern that was called recursively or as a subroutine contained
203        (*PRUNE) or any other control that caused it to give a non-standard return,
204        invalid errors such as "Error -26 (nested recursion at the same subject
205        position)" or even infinite loops could occur.
206    
207    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
208        computing the minimum length on reaching *ACCEPT, and so ended up with the
209        wrong value of 1 rather than 0. Further investigation indicates that
210        computing a minimum subject length in the presence of *ACCEPT is difficult
211        (think back references, subroutine calls), and so I have changed the code
212        so that no minimum is registered for a pattern that contains *ACCEPT.
213    
214    8.  If (*THEN) was present in the first (true) branch of a conditional group,
215        it was not handled as intended. [But see 16 below.]
216    
217    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
218        Sheri Pierce.
219    
220    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
221        the first byte in a match must be "a".
222    
223    11. Change 17 for 8.13 increased the recursion depth for patterns like
224        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
225        pattern contains any instances of (*THEN). If it does not, the old
226        optimizations are restored. It would be nice to do this on a per-group
227        basis, but at the moment that is not feasible.
228    
229    12. In some environments, the output of pcretest -C is CRLF terminated. This
230        broke RunTest's code that checks for the link size. A single white space
231        character after the value is now allowed for.
232    
233    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
234        For "fr", it uses the Windows-specific input and output files.
235    
236    14. If (*THEN) appeared in a group that was called recursively or as a
237        subroutine, it did not work as intended. [But see next item.]
238    
239    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
240        pattern fragments (but not containing any | characters). If A and B are
241        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
242        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
243        D. In other words, Perl considers parentheses that do not contain any |
244        characters to be part of a surrounding alternative, whereas PCRE was
245        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
246        differently. PCRE now behaves in the same way as Perl, except in the case
247        of subroutine/recursion calls such as (?1) which have in any case always
248        been different (but PCRE had them first :-).
249    
250    16. Related to 15 above: Perl does not treat the | in a conditional group as
251        creating alternatives. Such a group is treated in the same way as an
252        ordinary group without any | characters when processing (*THEN). PCRE has
253        been changed to match Perl's behaviour.
254    
255    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
256        RunGrepTest script failed.
257    
258    18. Change 22 for version 13 caused atomic groups to use more stack. This is
259        inevitable for groups that contain captures, but it can lead to a lot of
260        stack use in large patterns. The old behaviour has been restored for atomic
261        groups that do not contain any capturing parentheses.
262    
263    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
264        suppress the check for a minimum subject length at run time. (If it was
265        given to pcre_exec() or pcre_dfa_exec() it did work.)
266    
267    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
268        fail to work when decoding hex characters in data strings in EBCDIC
269        environments.
270    
271    21. It appears that in at least one Mac OS environment, the isxdigit() function
272        is implemented as a macro that evaluates to its argument more than once,
273        contravening the C 90 Standard (I haven't checked a later standard). There
274        was an instance in pcretest which caused it to go wrong when processing
275        \x{...} escapes in subject strings. The has been rewritten to avoid using
276        things like p++ in the argument of isxdigit().
277    
278    
279    Version 8.13 16-Aug-2011
280    ------------------------
281    
282    1.  The Unicode data tables have been updated to Unicode 6.0.0.
283    
284    2.  Two minor typos in pcre_internal.h have been fixed.
285    
286    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
287        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
288        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
289    
290    4.  There were a number of related bugs in the code for matching backrefences
291        caselessly in UTF-8 mode when codes for the characters concerned were
292        different numbers of bytes. For example, U+023A and U+2C65 are an upper
293        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
294        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
295        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
296        2-byte code at the end of the subject (it thought there wasn't enough data
297        left).
298    
299    5.  Comprehensive information about what went wrong is now returned by
300        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
301        as the output vector has at least 2 elements. The offset of the start of
302        the failing character and a reason code are placed in the vector.
303    
304    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
305        now returned is for the first byte of the failing character, instead of the
306        last byte inspected. This is an incompatible change, but I hope it is small
307        enough not to be a problem. It makes the returned offset consistent with
308        pcre_exec() and pcre_dfa_exec().
309    
310    7.  pcretest now gives a text phrase as well as the error number when
311        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
312        failure, the offset and reason code are output.
313    
314    8.  When \R was used with a maximizing quantifier it failed to skip backwards
315        over a \r\n pair if the subsequent match failed. Instead, it just skipped
316        back over a single character (\n). This seems wrong (because it treated the
317        two characters as a single entity when going forwards), conflicts with the
318        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
319        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
320        has been changed.
321    
322    9.  Some internal refactoring has changed the processing so that the handling
323        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
324        time (the PCRE_DOTALL option was changed this way some time ago: version
325        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
326        which was always a bit of a fudge. It also means that there is one less
327        argument for the match() function, which reduces its stack requirements
328        slightly. This change also fixes an incompatibility with Perl: the pattern
329        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
330    
331    10. More internal refactoring has drastically reduced the number of recursive
332        calls to match() for possessively repeated groups such as (abc)++ when
333        using pcre_exec().
334    
335    11. While implementing 10, a number of bugs in the handling of groups were
336        discovered and fixed:
337    
338        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
339        (a|)*(?1) gave a compile-time internal error.
340        ((a|)+)+  did not notice that the outer group could match an empty string.
341        (^a|^)+   was not marked as anchored.
342        (.*a|.*)+ was not marked as matching at start or after a newline.
343    
344    12. Yet more internal refactoring has removed another argument from the match()
345        function. Special calls to this function are now indicated by setting a
346        value in a variable in the "match data" data block.
347    
348    13. Be more explicit in pcre_study() instead of relying on "default" for
349        opcodes that mean there is no starting character; this means that when new
350        ones are added and accidentally left out of pcre_study(), testing should
351        pick them up.
352    
353    14. The -s option of pcretest has been documented for ages as being an old
354        synonym of -m (show memory usage). I have changed it to mean "force study
355        for every regex", that is, assume /S for every regex. This is similar to -i
356        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
357        using it. It makes it easier to run collections of tests with and without
358        study enabled, and thereby test pcre_study() more easily. All the standard
359        tests are now run with and without -s (but some patterns can be marked as
360        "never study" - see 20 below).
361    
362    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
363        restoration of the capturing data to the outer values was not happening
364        correctly.
365    
366    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
367        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
368        pattern had matched an empty string, and so incorrectly returned a no
369        match.
370    
371    17. There was optimizing code for the last branch of non-capturing parentheses,
372        and also for the obeyed branch of a conditional subexpression, which used
373        tail recursion to cut down on stack usage. Unfortunately, now that there is
374        the possibility of (*THEN) occurring in these branches, tail recursion is
375        no longer possible because the return has to be checked for (*THEN). These
376        two optimizations have therefore been removed. [But see 8.20/11 above.]
377    
378    18. If a pattern containing \R was studied, it was assumed that \R always
379        matched two bytes, thus causing the minimum subject length to be
380        incorrectly computed because \R can also match just one byte.
381    
382    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
383        was incorrectly computed.
384    
385    20. If /S is present twice on a test pattern in pcretest input, it now
386        *disables* studying, thereby overriding the use of -s on the command line
387        (see 14 above). This is necessary for one or two tests to keep the output
388        identical in both cases.
389    
390    21. When (*ACCEPT) was used in an assertion that matched an empty string and
391        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
392    
393    22. When an atomic group that contained a capturing parenthesis was
394        successfully matched, but the branch in which it appeared failed, the
395        capturing was not being forgotten if a higher numbered group was later
396        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
397        group 1 to "a", when in fact it should be unset. This applied to multi-
398        branched capturing and non-capturing groups, repeated or not, and also to
399        positive assertions (capturing in negative assertions does not happen
400        in PCRE) and also to nested atomic groups.
401    
402    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
403        subject after a captured substring, to make it easier to tell which of a
404        number of identical substrings has been captured.
405    
406    24. The way atomic groups are processed by pcre_exec() has been changed so that
407        if they are repeated, backtracking one repetition now resets captured
408        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
409        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
410        "aaa". Previously, it would have been "a". As part of this code
411        refactoring, the way recursive calls are handled has also been changed.
412    
413    25. If an assertion condition captured any substrings, they were not passed
414        back unless some other capturing happened later. For example, if
415        (?(?=(a))a) was matched against "a", no capturing was returned.
416    
417    26. When studying a pattern that contained subroutine calls or assertions,
418        the code for finding the minimum length of a possible match was handling
419        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
420        group 1 called group 2 while simultaneously a separate group 2 called group
421        1). A stack overflow occurred in this case. I have fixed this by limiting
422        the recursion depth to 10.
423    
424    27. Updated RunTest.bat in the distribution to the version supplied by Tom
425        Fortmann. This supports explicit test numbers on the command line, and has
426        argument validation and error reporting.
427    
428    28. An instance of \X with an unlimited repeat could fail if at any point the
429        first character it looked at was a mark character.
430    
431    29. Some minor code refactoring concerning Unicode properties and scripts
432        should reduce the stack requirement of match() slightly.
433    
434    30. Added the '=' option to pcretest to check the setting of unused capturing
435        slots at the end of the pattern, which are documented as being -1, but are
436        not included in the return count.
437    
438    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
439        compiled something random. Now it gives a compile-time error (as does
440        Perl).
441    
442    32. A *MARK encountered during the processing of a positive assertion is now
443        recorded and passed back (compatible with Perl).
444    
445    33. If --only-matching or --colour was set on a pcregrep call whose pattern
446        had alternative anchored branches, the search for a second match in a line
447        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
448        matched the line "0102" twice. The same bug affected patterns that started
449        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
450        twice.
451    
452    34. Previously, PCRE did not allow quantification of assertions. However, Perl
453        does, and because of capturing effects, quantifying parenthesized
454        assertions may at times be useful. Quantifiers are now allowed for
455        parenthesized assertions.
456    
457    35. A minor code tidy in pcre_compile() when checking options for \R usage.
458    
459    36. \g was being checked for fancy things in a character class, when it should
460        just be a literal "g".
461    
462    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
463        appearance of a nested POSIX class supersedes an apparent external class.
464        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
465        unescaped square brackets may also appear as part of class names. For
466        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
467        more like Perl. (But see 8.20/1 above.)
468    
469    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
470        was because it thought it was \N{name}, which is not supported).
471    
472    39. Add minix to OS list not supporting the -S option in pcretest.
473    
474    40. PCRE tries to detect cases of infinite recursion at compile time, but it
475        cannot analyze patterns in sufficient detail to catch mutual recursions
476        such as ((?1))((?2)). There is now a runtime test that gives an error if a
477        subgroup is called recursively as a subpattern for a second time at the
478        same position in the subject string. In previous releases this might have
479        been caught by the recursion limit, or it might have run out of stack.
480    
481    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
482        happen only once. PCRE was, however incorrectly giving a compile time error
483        "recursive call could loop indefinitely" because it cannot analyze the
484        pattern in sufficient detail. The compile time test no longer happens when
485        PCRE is compiling a conditional subpattern, but actual runaway loops are
486        now caught at runtime (see 40 above).
487    
488    42. It seems that Perl allows any characters other than a closing parenthesis
489        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
490        has been changed to be the same.
491    
492    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
493        as not to get warnings when autogen.sh is called. Also changed
494        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
495    
496    44. To help people who use pcregrep to scan files containing exceedingly long
497        lines, the following changes have been made:
498    
499        (a) The default value of the buffer size parameter has been increased from
500            8K to 20K. (The actual buffer used is three times this size.)
501    
502        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
503            PCRE is built.
504    
505        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
506            to be set at run time.
507    
508        (d) Numerical values in pcregrep options can be followed by K or M, for
509            example --buffer-size=50K.
510    
511        (e) If a line being scanned overflows pcregrep's buffer, an error is now
512            given and the return code is set to 2.
513    
514    45. Add a pointer to the latest mark to the callout data block.
515    
516    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
517        partial match of an empty string instead of no match. This was specific to
518        the use of ".".
519    
520    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
521        complete match instead of a partial match. This bug was dependent on both
522        the PCRE_UTF8 and PCRE_DOTALL options being set.
523    
524    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
525        starting byte set, because \b was not being ignored.
526    
527    
528    Version 8.12 15-Jan-2011
529    ------------------------
530    
531    1.  Fixed some typos in the markup of the man pages, and wrote a script that
532        checks for such things as part of the documentation building process.
533    
534    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
535        --match-limit and --recursion-limit options (added for 8.11). In
536        particular, this made one of the standard tests fail. (The integer value
537        went into the wrong half of a long int.)
538    
539    3.  If the --colour option was given to pcregrep with -v (invert match), it
540        did strange things, either producing crazy output, or crashing. It should,
541        of course, ignore a request for colour when reporting lines that do not
542        match.
543    
544    4.  Another pcregrep bug caused similar problems if --colour was specified with
545        -M (multiline) and the pattern match finished with a line ending.
546    
547    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
548        matched in multiline mode, the following line was shown as part of the
549        match. This seems wrong, so I have changed it.
550    
551    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
552        the check for further matches in the same line (so they could be coloured)
553        to overrun the end of the current line. If another match was found, it was
554        incorrectly shown (and then shown again when found in the next line).
555    
556    7.  If pcregrep was compiled under Windows, there was a reference to the
557        function pcregrep_exit() before it was defined. I am assuming this was
558        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
559        reported by a user. I've moved the definition above the reference.
560    
561    
562    Version 8.11 10-Dec-2010
563    ------------------------
564    
565    1.  (*THEN) was not working properly if there were untried alternatives prior
566        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
567        backtracked to try for "b" instead of moving to the next alternative branch
568        at the same level (in this case, to look for "c"). The Perl documentation
569        is clear that when (*THEN) is backtracked onto, it goes to the "next
570        alternative in the innermost enclosing group".
571    
572    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
573        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
574        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
575        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
576        (*THEN).
577    
578    3.  If \s appeared in a character class, it removed the VT character from
579        the class, even if it had been included by some previous item, for example
580        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
581        of \s, but is part of the POSIX "space" class.)
582    
583    4.  A partial match never returns an empty string (because you can always
584        match an empty string at the end of the subject); however the checking for
585        an empty string was starting at the "start of match" point. This has been
586        changed to the "earliest inspected character" point, because the returned
587        data for a partial match starts at this character. This means that, for
588        example, /(?<=abc)def/ gives a partial match for the subject "abc"
589        (previously it gave "no match").
590    
591    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
592        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
593        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
594        has an implication that the given string is incomplete (because a partial
595        match is preferred over a full match). For this reason, these items now
596        give a partial match in this situation. [Aside: previously, the one case
597        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
598        match rather than a full match, which was wrong by the old rules, but is
599        now correct.]
600    
601    6.  There was a bug in the handling of #-introduced comments, recognized when
602        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
603        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
604        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
605        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
606        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
607        places in pcre_compile().
608    
609    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
610        comments when looking ahead for named forward references to subpatterns,
611        the only newline sequence it recognized was NL. It now handles newlines
612        according to the set newline convention.
613    
614    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
615        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
616        cater for a lack of strerror(). These oversights have been fixed.
617    
618    9.  Added --match-limit and --recursion-limit to pcregrep.
619    
620    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
621    
622    11. When the -o option was used, pcregrep was setting a return code of 1, even
623        when matches were found, and --line-buffered was not being honoured.
624    
625    12. Added an optional parentheses number to the -o and --only-matching options
626        of pcregrep.
627    
628    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
629        can match an empty string. The code to do it in pcretest and pcredemo
630        needed fixing:
631    
632        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
633            only one byte after an empty string match just before CRLF (this case
634            just got forgotten; "any" and "anycrlf" were OK).
635    
636        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
637            mode when an empty string match preceded an ASCII character followed by
638            a non-ASCII character. (The code for advancing by one character rather
639            than one byte was nonsense.)
640    
641        (c) The pcredemo.c sample program did not have any code at all to handle
642            the cases when CRLF is a valid newline sequence.
643    
644    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
645        as a starting offset was within the subject string. There is now a new
646        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
647        negative or greater than the length of the string. In order to test this,
648        pcretest is extended to allow the setting of negative starting offsets.
649    
650    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
651        starting offset points to the beginning of a UTF-8 character was
652        unnecessarily clumsy. I tidied it up.
653    
654    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
655        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
656    
657    17. Nobody had reported that the --include_dir option, which was added in
658        release 7.7 should have been called --include-dir (hyphen, not underscore)
659        for compatibility with GNU grep. I have changed it to --include-dir, but
660        left --include_dir as an undocumented synonym, and the same for
661        --exclude-dir, though that is not available in GNU grep, at least as of
662        release 2.5.4.
663    
664    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
665        characters from a string of bytes) have been redefined so as not to use
666        loops, in order to improve performance in some environments. At the same
667        time, I abstracted some of the common code into auxiliary macros to save
668        repetition (this should not affect the compiled code).
669    
670    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
671        compile-time error is now given if \c is not followed by an ASCII
672        character, that is, a byte less than 128. (In EBCDIC mode, the code is
673        different, and any byte value is allowed.)
674    
675    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
676        START_OPTIMIZE option, which is now allowed at compile time - but just
677        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
678        to pcregrep and other applications that have no direct access to PCRE
679        options. The new /Y option in pcretest sets this option when calling
680        pcre_compile().
681    
682    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
683        back references. Groups containing recursive back references were forced to
684        be atomic by that change, but in the case of named groups, the amount of
685        memory required was incorrectly computed, leading to "Failed: internal
686        error: code overflow". This has been fixed.
687    
688    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
689        pcretest.c, to avoid build problems in some Borland environments.
690    
691    
692    Version 8.10 25-Jun-2010
693    ------------------------
694    
695    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
696        THEN.
697    
698    2.  (*ACCEPT) was not working when inside an atomic group.
699    
700    3.  Inside a character class, \B is treated as a literal by default, but
701        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
702        causes the error). The code is unchanged, but I tidied the documentation.
703    
704    4.  Inside a character class, PCRE always treated \R and \X as literals,
705        whereas Perl faults them if its -w option is set. I have changed PCRE so
706        that it faults them when PCRE_EXTRA is set.
707    
708    5.  Added support for \N, which always matches any character other than
709        newline. (It is the same as "." when PCRE_DOTALL is not set.)
710    
711    6.  When compiling pcregrep with newer versions of gcc which may have
712        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
713        declared with attribute warn_unused_result" were given. Just casting the
714        result to (void) does not stop the warnings; a more elaborate fudge is
715        needed. I've used a macro to implement this.
716    
717    7.  Minor change to pcretest.c to avoid a compiler warning.
718    
719    8.  Added four artifical Unicode properties to help with an option to make
720        \s etc use properties (see next item). The new properties are: Xan
721        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
722    
723    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
724        use Unicode properties. (*UCP) at the start of a pattern can be used to set
725        this option. Modified pcretest to add /W to test this facility. Added
726        REG_UCP to make it available via the POSIX interface.
727    
728    10. Added --line-buffered to pcregrep.
729    
730    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
731        studied, and the match started with a letter with a code point greater than
732        127 whose first byte was different to the first byte of the other case of
733        the letter, the other case of this starting letter was not recognized
734        (#976).
735    
736    12. If a pattern that was studied started with a repeated Unicode property
737        test, for example, \p{Nd}+, there was the theoretical possibility of
738        setting up an incorrect bitmap of starting bytes, but fortunately it could
739        not have actually happened in practice until change 8 above was made (it
740        added property types that matched character-matching opcodes).
741    
742    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
743        possible starting bytes for non-anchored patterns.
744    
745    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
746        \R, and also a number of cases that involve Unicode properties, both
747        explicit and implicit when PCRE_UCP is set.
748    
749    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
750        input, it could crash or give wrong results if characters with values
751        greater than 0xc0 were present in the subject string. (Detail: it assumed
752        UTF-8 input when processing these items.)
753    
754    16. Added a lot of (int) casts to avoid compiler warnings in systems where
755        size_t is 64-bit (#991).
756    
757    17. Added a check for running out of memory when PCRE is compiled with
758        --disable-stack-for-recursion (#990).
759    
760    18. If the last data line in a file for pcretest does not have a newline on
761        the end, a newline was missing in the output.
762    
763    19. The default pcre_chartables.c file recognizes only ASCII characters (values
764        less than 128) in its various bitmaps. However, there is a facility for
765        generating tables according to the current locale when PCRE is compiled. It
766        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
767        characters, are recognized by isspace() and therefore were getting set in
768        these tables, and indeed these tables seem to approximate to ISO 8859. This
769        caused a problem in UTF-8 mode when pcre_study() was used to create a list
770        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
771        which of course cannot start UTF-8 characters. I have changed the code so
772        that only real ASCII characters (less than 128) and the correct starting
773        bytes for UTF-8 encodings are set for characters greater than 127 when in
774        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
775        altogether.)
776    
777    20. Added the /T option to pcretest so as to be able to run tests with non-
778        standard character tables, thus making it possible to include the tests
779        used for 19 above in the standard set of tests.
780    
781    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
782        reference to a subpattern the other side of a comment that contains an
783        opening parenthesis caused either an internal compiling error, or a
784        reference to the wrong subpattern.
785    
786    
787    Version 8.02 19-Mar-2010
788    ------------------------
789    
790    1.  The Unicode data tables have been updated to Unicode 5.2.0.
791    
792    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
793        configured.
794    
795    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
796        original author of that file, following a query about its status.
797    
798    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
799        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
800    
801    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
802        quantifier applied to a forward-referencing subroutine call, could compile
803        incorrect code or give the error "internal error: previously-checked
804        referenced subpattern not found".
805    
806    6.  Both MS Visual Studio and Symbian OS have problems with initializing
807        variables to point to external functions. For these systems, therefore,
808        pcre_malloc etc. are now initialized to local functions that call the
809        relevant global functions.
810    
811    7.  There were two entries missing in the vectors called coptable and poptable
812        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
813        I've fixed the data, and added a kludgy way of testing at compile time that
814        the lengths are correct (equal to the number of opcodes).
815    
816    8.  Following on from 7, I added a similar kludge to check the length of the
817        eint vector in pcreposix.c.
818    
819    9.  Error texts for pcre_compile() are held as one long string to avoid too
820        much relocation at load time. To find a text, the string is searched,
821        counting zeros. There was no check for running off the end of the string,
822        which could happen if a new error number was added without updating the
823        string.
824    
825    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
826    
827    11. \K was not working if it appeared in an atomic group or in a group that
828        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
829        \K is "not well defined" if used in an assertion. PCRE now accepts it if
830        the assertion is positive, but not if it is negative.
831    
832    12. Change 11 fortuitously reduced the size of the stack frame used in the
833        "match()" function of pcre_exec.c by one pointer. Forthcoming
834        implementation of support for (*MARK) will need an extra pointer on the
835        stack; I have reserved it now, so that the stack frame size does not
836        decrease.
837    
838    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
839        item in branch that calls a recursion is a subroutine call - as in the
840        second branch in the above example - was incorrectly given the compile-
841        time error "recursive call could loop indefinitely" because pcre_compile()
842        was not correctly checking the subroutine for matching a non-empty string.
843    
844    14. The checks for overrunning compiling workspace could trigger after an
845        overrun had occurred. This is a "should never occur" error, but it can be
846        triggered by pathological patterns such as hundreds of nested parentheses.
847        The checks now trigger 100 bytes before the end of the workspace.
848    
849    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
850    
851    
852    Version 8.01 19-Jan-2010
853    ------------------------
854    
855    1.  If a pattern contained a conditional subpattern with only one branch (in
856        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
857        computed the wrong minimum data length (which is of course zero for such
858        subpatterns). This could cause incorrect "no match" results.
859    
860    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
861        the pattern is reset in the first branch, pcre_compile() failed with
862        "internal error: code overflow at offset...". This happened only when
863        the reset was to the original external option setting. (An optimization
864        abstracts leading options settings into an external setting, which was the
865        cause of this.)
866    
867    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
868        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
869        assertion pattern did not match (meaning that the assertion was true), it
870        was incorrectly treated as false if the SKIP had been reached during the
871        matching. This also applied to assertions used as conditions.
872    
873    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
874        assertion subpattern, including such a pattern used as a condition,
875        unpredictable results occurred, instead of the error return
876        PCRE_ERROR_DFA_UITEM.
877    
878    5.  The C++ GlobalReplace function was not working like Perl for the special
879        situation when an empty string is matched. It now does the fancy magic
880        stuff that is necessary.
881    
882    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
883        removed. (These were left over from very, very early versions of PCRE.)
884    
885    7.  Some cosmetic changes to the code to make life easier when compiling it
886        as part of something else:
887    
888        (a) Change DEBUG to PCRE_DEBUG.
889    
890        (b) In pcre_compile(), rename the member of the "branch_chain" structure
891            called "current" as "current_branch", to prevent a collision with the
892            Linux macro when compiled as a kernel module.
893    
894        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
895            prevent a collision with the Linux macro when compiled as a kernel
896            module.
897    
898    8.  In pcre_compile() there are some checks for integer overflows that used to
899        cast potentially large values to (double). This has been changed to that
900        when building, a check for int64_t is made, and if it is found, it is used
901        instead, thus avoiding the use of floating point arithmetic. (There is no
902        other use of FP in PCRE.) If int64_t is not found, the fallback is to
903        double.
904    
905    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
906        2005 (difference between two addresses compared to an unsigned value).
907    
908    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
909        custom one, because of the following reported problem in Windows:
910    
911          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
912              under Win32.
913          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
914              therefore missing the function definition.
915          - The compiler thus generates a "C" signature for the test function.
916          - The linker fails to find the "C" function.
917          - PCRE fails to configure if asked to do so against libbz2.
918    
919    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
920        messages were output:
921    
922          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
923          rerunning libtoolize, to keep the correct libtool macros in-tree.
924          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
925    
926        I have done both of these things.
927    
928    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
929        most of the time, it *can* run out if it is given a pattern that contains a
930        runaway infinite recursion. I updated the discussion in the pcrestack man
931        page.
932    
933    13. Now that we have gone to the x.xx style of version numbers, the minor
934        version may start with zero. Using 08 or 09 is a bad idea because users
935        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
936        interpreted as invalid octal numbers. I've updated the previous comment in
937        configure.ac, and also added a check that gives an error if 08 or 09 are
938        used.
939    
940    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
941        causing partial matching to fail when the end of the subject matched \W
942        in a UTF-8 pattern where \W was quantified with a minimum of 3.
943    
944    15. There were some discrepancies between the declarations in pcre_internal.h
945        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
946        their definitions. The declarations used "const uschar *" and the
947        definitions used USPTR. Even though USPTR is normally defined as "const
948        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
949        reported that: "This difference in casting confuses some C++ compilers, for
950        example, SunCC recognizes above declarations as different functions and
951        generates broken code for hbpcre." I have changed the declarations to use
952        USPTR.
953    
954    16. GNU libtool is named differently on some systems. The autogen.sh script now
955        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
956        (FreeBSD).
957    
958    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
959        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
960        comment: "Figure out how to create a longlong from a string: strtoll and
961        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
962        instance, but it only takes 2 args instead of 3!"
963    
964    18. A subtle bug concerned with back references has been fixed by a change of
965        specification, with a corresponding code fix. A pattern such as
966        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
967        refers, was giving matches when it shouldn't. For example, xa=xaaa would
968        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
969        same bug. Such groups have to be quantified to be useful, or contained
970        inside another quantified group. (If there's no repetition, the reference
971        can never match.) The problem arises because, having left the group and
972        moved on to the rest of the pattern, a later failure that backtracks into
973        the group uses the captured value from the final iteration of the group
974        rather than the correct earlier one. I have fixed this in PCRE by forcing
975        any group that contains a reference to itself to be an atomic group; that
976        is, there cannot be any backtracking into it once it has completed. This is
977        similar to recursive and subroutine calls.
978    
979    
980    Version 8.00 19-Oct-09
981    ----------------------
982    
983    1.  The table for translating pcre_compile() error codes into POSIX error codes
984        was out-of-date, and there was no check on the pcre_compile() error code
985        being within the table. This could lead to an OK return being given in
986        error.
987    
988    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
989        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
990        in a Windows environment.
991    
992    3.  The pcregrep --count option prints the count for each file even when it is
993        zero, as does GNU grep. However, pcregrep was also printing all files when
994        --files-with-matches was added. Now, when both options are given, it prints
995        counts only for those files that have at least one match. (GNU grep just
996        prints the file name in this circumstance, but including the count seems
997        more useful - otherwise, why use --count?) Also ensured that the
998        combination -clh just lists non-zero counts, with no names.
999    
1000    4.  The long form of the pcregrep -F option was incorrectly implemented as
1001        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1002        but it seems right to fix it, and I didn't think it was worth preserving
1003        the old behaviour.
1004    
1005    5.  The command line items --regex=pattern and --regexp=pattern were not
1006        recognized by pcregrep, which required --regex pattern or --regexp pattern
1007        (with a space rather than an '='). The man page documented the '=' forms,
1008        which are compatible with GNU grep; these now work.
1009    
1010    6.  No libpcreposix.pc file was created for pkg-config; there was just
1011        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1012    
1013    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1014        when UCP support is not needed, by modifying the Python script that
1015        generates it from Unicode data files. This should not matter if the module
1016        is correctly used as a library, but I received one complaint about 50K of
1017        unwanted data. My guess is that the person linked everything into his
1018        program rather than using a library. Anyway, it does no harm.
1019    
1020    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1021        was a minimum greater than 1 for a wide character in a possessive
1022        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1023        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1024        character. Chaos in the form of incorrect output or a compiling loop could
1025        result.
1026    
1027    9.  The restrictions on what a pattern can contain when partial matching is
1028        requested for pcre_exec() have been removed. All patterns can now be
1029        partially matched by this function. In addition, if there are at least two
1030        slots in the offset vector, the offset of the earliest inspected character
1031        for the match and the offset of the end of the subject are set in them when
1032        PCRE_ERROR_PARTIAL is returned.
1033    
1034    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1035        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1036        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1037        and may be more useful for multi-segment matching.
1038    
1039    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1040        used to be given if ever the end of the subject was reached; now it is
1041        given only if matching could not proceed because another character was
1042        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1043        string "Z", which now yields "no match" instead of "partial match". In the
1044        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1045        final character ended with (*FAIL).
1046    
1047    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1048        if the pattern had a "must contain" character that was already found in the
1049        earlier partial match, unless partial matching was again requested. For
1050        example, with the pattern /dog.(body)?/, the "must contain" character is
1051        "g". If the first part-match was for the string "dog", restarting with
1052        "sbody" failed. This bug has been fixed.
1053    
1054    13. The string returned by pcre_dfa_exec() after a partial match has been
1055        changed so that it starts at the first inspected character rather than the
1056        first character of the match. This makes a difference only if the pattern
1057        starts with a lookbehind assertion or \b or \B (\K is not supported by
1058        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1059        matching functions compatible, and I think it's the right thing to do.
1060    
1061    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1062        so that the demonstration program is easily available in environments where
1063        PCRE has not been installed from source.
1064    
1065    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1066        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1067        library.
1068    
1069    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1070        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1071        is not the first non-POSIX option to be added. Clearly some people find
1072        these options useful.
1073    
1074    17. If a caller to the POSIX matching function regexec() passes a non-zero
1075        value for nmatch with a NULL value for pmatch, the value of
1076        nmatch is forced to zero.
1077    
1078    18. RunGrepTest did not have a test for the availability of the -u option of
1079        the diff command, as RunTest does. It now checks in the same way as
1080        RunTest, and also checks for the -b option.
1081    
1082    19. If an odd number of negated classes containing just a single character
1083        interposed, within parentheses, between a forward reference to a named
1084        subpattern and the definition of the subpattern, compilation crashed with
1085        an internal error, complaining that it could not find the referenced
1086        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1087        [The bug was that it was starting one character too far in when skipping
1088        over the character class, thus treating the ] as data rather than
1089        terminating the class. This meant it could skip too much.]
1090    
1091    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1092        /g option in pcretest when the pattern contains \K, which makes it possible
1093        to have an empty string match not at the start, even when the pattern is
1094        anchored. Updated pcretest and pcredemo to use this option.
1095    
1096    21. If the maximum number of capturing subpatterns in a recursion was greater
1097        than the maximum at the outer level, the higher number was returned, but
1098        with unset values at the outer level. The correct (outer level) value is
1099        now given.
1100    
1101    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1102        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1103        make it do so. The string so far is captured, making this feature
1104        compatible with Perl.
1105    
1106    23. The tests have been re-organized, adding tests 11 and 12, to make it
1107        possible to check the Perl 5.10 features against Perl 5.10.
1108    
1109    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1110        pattern matches a fixed length string. PCRE did not allow this; now it
1111        does. Neither allows recursion.
1112    
1113    25. I finally figured out how to implement a request to provide the minimum
1114        length of subject string that was needed in order to match a given pattern.
1115        (It was back references and recursion that I had previously got hung up
1116        on.) This code has now been added to pcre_study(); it finds a lower bound
1117        to the length of subject needed. It is not necessarily the greatest lower
1118        bound, but using it to avoid searching strings that are too short does give
1119        some useful speed-ups. The value is available to calling programs via
1120        pcre_fullinfo().
1121    
1122    26. While implementing 25, I discovered to my embarrassment that pcretest had
1123        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1124        study optimizations had never been tested with that matching function.
1125        Oops. What is worse, even when it was passed study data, there was a bug in
1126        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1127        were also very few tests of studied patterns with pcre_dfa_exec().
1128    
1129    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1130        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1131        on the other side of the coin, they are no longer allowed to have different
1132        names, because these cannot be distinguished in PCRE, and this has caused
1133        confusion. (This is a difference from Perl.)
1134    
1135    28. When duplicate subpattern names are present (necessarily with different
1136        numbers, as required by 27 above), and a test is made by name in a
1137        conditional pattern, either for a subpattern having been matched, or for
1138        recursion in such a pattern, all the associated numbered subpatterns are
1139        tested, and the overall condition is true if the condition is true for any
1140        one of them. This is the way Perl works, and is also more like the way
1141        testing by number works.
1142    
1143    
1144    Version 7.9 11-Apr-09
1145    ---------------------
1146    
1147    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1148        (pcretest), all targets were linked against these libraries. This included
1149        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1150        libraries. This caused unwanted dependencies to be created. This problem
1151        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1152        pcretest is linked with readline.
1153    
1154    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1155        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1156        moved outside it again, because FALSE and TRUE are already defined in AIX,
1157        but BOOL is not.
1158    
1159    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1160        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1161    
1162    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1163        hyphens) following file names and line numbers when outputting matching
1164        lines. This is not true; no spaces are inserted. I have also clarified the
1165        wording for the --colour (or --color) option.
1166    
1167    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1168        was not coloured; this is different to GNU grep, so I have changed it to be
1169        the same.
1170    
1171    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1172        each matching line was coloured. Now it goes on to look for further matches
1173        of any of the test patterns, which is the same behaviour as GNU grep.
1174    
1175    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1176        doesn't make sense to accept an empty string match in pcregrep, so I have
1177        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1178        seems to be how GNU grep behaves.
1179    
1180    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1181        start or after a newline", because the conditional assertion was not being
1182        correctly handled. The rule now is that both the assertion and what follows
1183        in the first alternative must satisfy the test.
1184    
1185    9.  If auto-callout was enabled in a pattern with a conditional group whose
1186        condition was an assertion, PCRE could crash during matching, both with
1187        pcre_exec() and pcre_dfa_exec().
1188    
1189    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1190        used for matching.
1191    
1192    11. Unicode property support in character classes was not working for
1193        characters (bytes) greater than 127 when not in UTF-8 mode.
1194    
1195    12. Added the -M command line option to pcretest.
1196    
1197    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1198    
1199    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1200    
1201    16. Added comments and documentation about mis-use of no_arg in the C++
1202        wrapper.
1203    
1204    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1205        from Martin Jerabek that uses macro names for all relevant character and
1206        string constants.
1207    
1208    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1209        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1210        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1211        these, but not everybody uses configure.
1212    
1213    19. A conditional group that had only one branch was not being correctly
1214        recognized as an item that could match an empty string. This meant that an
1215        enclosing group might also not be so recognized, causing infinite looping
1216        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1217        with the subject "ab", where knowledge that the repeated group can match
1218        nothing is needed in order to break the loop.
1219    
1220    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1221        exec(), but without supplying a callout function, matching went wrong.
1222    
1223    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1224        leak if the size of the offset vector was greater than 30. When the vector
1225        is smaller, the saved offsets during recursion go onto a local stack
1226        vector, but for larger vectors malloc() is used. It was failing to free
1227        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1228        error, in fact).
1229    
1230    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1231        heapframe that is used only when UTF-8 support is enabled. This caused no
1232        problem, but was untidy.
1233    
1234    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1235        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1236        included within another project.
1237    
1238    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1239        slightly modified by me:
1240    
1241          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1242              not building pcregrep.
1243    
1244          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1245              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1246    
1247    25. Forward references, both numeric and by name, in patterns that made use of
1248        duplicate group numbers, could behave incorrectly or give incorrect errors,
1249        because when scanning forward to find the reference group, PCRE was not
1250        taking into account the duplicate group numbers. A pattern such as
1251        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1252    
1253    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1254        the feature of a custom pointer more persuasive (as requested by a user).
1255    
1256    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1257        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1258        pre-defined.
1259    
1260    28. Added support for (*UTF8) at the start of a pattern.
1261    
1262    29. Arrange for flags added by the "release type" setting in CMake to be shown
1263        in the configuration summary.
1264    
1265    
1266    Version 7.8 05-Sep-08
1267    ---------------------
1268    
1269    1.  Replaced UCP searching code with optimized version as implemented for Ad
1270        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1271        stage table and inline lookup instead of a function, giving speed ups of 2
1272        to 5 times on some simple patterns that I tested. Permission was given to
1273        distribute the MultiStage2.py script that generates the tables (it's not in
1274        the tarball, but is in the Subversion repository).
1275    
1276    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1277        scripts.
1278    
1279    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1280        a group with a zero qualifier. The result of the study could be incorrect,
1281        or the function might crash, depending on the pattern.
1282    
1283    4.  Caseless matching was not working for non-ASCII characters in back
1284        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1285        It now works when Unicode Property Support is available.
1286    
1287    5.  In pcretest, an escape such as \x{de} in the data was always generating
1288        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1289        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1290        truncation.
1291    
1292    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1293    
1294    7.  Added two (int) casts to pcregrep when printing the difference of two
1295        pointers, in case they are 64-bit values.
1296    
1297    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1298        test 2 if it fails.
1299    
1300    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1301        and a #define of that name to empty if it is not externally set. This is to
1302        allow users of MSVC to set it if necessary.
1303    
1304    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1305        the convenience functions in the pcre_get.c source file.
1306    
1307    11. An option change at the start of a pattern that had top-level alternatives
1308        could cause overwriting and/or a crash. This command provoked a crash in
1309        some environments:
1310    
1311          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1312    
1313        This potential security problem was recorded as CVE-2008-2371.
1314    
1315    12. For a pattern where the match had to start at the beginning or immediately
1316        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1317        pcre_dfa_exec() could read past the end of the passed subject if there was
1318        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1319        pcretest so that it places the subject at the end of its malloc-ed buffer.
1320    
1321    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1322        exec() might read past the end of the data buffer in UTF-8 mode.
1323    
1324    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1325        the data contained the byte 0x85 as part of a UTF-8 character within its
1326        first line. This applied both to normal and DFA matching.
1327    
1328    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1329        /^[^d]*?$/8 failed to match "abc".
1330    
1331    16. Added a missing copyright notice to pcrecpp_internal.h.
1332    
1333    17. Make it more clear in the documentation that values returned from
1334        pcre_exec() in ovector are byte offsets, not character counts.
1335    
1336    18. Tidied a few places to stop certain compilers from issuing warnings.
1337    
1338    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1339        supplied by Stefan Weber. I made a further small update for 7.8 because
1340        there is a change of source arrangements: the pcre_searchfuncs.c module is
1341        replaced by pcre_ucd.c.
1342    
1343    
1344    Version 7.7 07-May-08
1345    ---------------------
1346    
1347    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1348        a string to a long long, pretend we don't even have a long long." This is
1349        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1350    
1351    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1352        pre-7.6 versions, which defined a global no_arg variable instead of putting
1353        it in the RE class. (See also #8 below.)
1354    
1355    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1356        Lopes.
1357    
1358    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1359    
1360        (1) The include/exclude patterns were being applied to the whole pathnames
1361            of files, instead of just to the final components.
1362    
1363        (2) If there was more than one level of directory, the subdirectories were
1364            skipped unless they satisfied the include/exclude conditions. This is
1365            inconsistent with GNU grep (and could even be seen as contrary to the
1366            pcregrep specification - which I improved to make it absolutely clear).
1367            The action now is always to scan all levels of directory, and just
1368            apply the include/exclude patterns to regular files.
1369    
1370    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1371        --exclude_dir in the tests to avoid scanning .svn directories.
1372    
1373    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1374        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1375        doesn't support NULs in patterns.
1376    
1377    7.  Added some missing "const"s to declarations of static tables in
1378        pcre_compile.c and pcre_dfa_exec.c.
1379    
1380    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1381        caused by fix #2  above. (Subsequently also a second patch to fix the
1382        first patch. And a third patch - this was a messy problem.)
1383    
1384    9.  Applied Craig's patch to remove the use of push_back().
1385    
1386    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1387        matching function regexec().
1388    
1389    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1390        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1391        references. PCRE supports relative numbers with this syntax (I don't think
1392        Oniguruma does).
1393    
1394    12. Previously, a group with a zero repeat such as (...){0} was completely
1395        omitted from the compiled regex. However, this means that if the group
1396        was called as a subroutine from elsewhere in the pattern, things went wrong
1397        (an internal error was given). Such groups are now left in the compiled
1398        pattern, with a new opcode that causes them to be skipped at execution
1399        time.
1400    
1401    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1402        to the way PCRE behaves:
1403    
1404        (a) A lone ] character is dis-allowed (Perl treats it as data).
1405    
1406        (b) A back reference to an unmatched subpattern matches an empty string
1407            (Perl fails the current match path).
1408    
1409        (c) A data ] in a character class must be notated as \] because if the
1410            first data character in a class is ], it defines an empty class. (In
1411            Perl it is not possible to have an empty class.) The empty class []
1412            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1413            The negative empty class [^] matches any one character, independently
1414            of the DOTALL setting.
1415    
1416    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1417        non-existent subpattern following a character class starting with ']' and
1418        containing () gave an internal compiling error instead of "reference to
1419        non-existent subpattern". Fortunately, when the pattern did exist, the
1420        compiled code was correct. (When scanning forwards to check for the
1421        existencd of the subpattern, it was treating the data ']' as terminating
1422        the class, so got the count wrong. When actually compiling, the reference
1423        was subsequently set up correctly.)
1424    
1425    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1426        it was being rejected as not supported by pcre_dfa_exec(), even though
1427        other assertions are supported. I have made pcre_dfa_exec() support
1428        (*FAIL).
1429    
1430    16. The implementation of 13c above involved the invention of a new opcode,
1431        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1432        cannot be changed at match time, I realized I could make a small
1433        improvement to matching performance by compiling OP_ALLANY instead of
1434        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1435        on the OP_ANY path.
1436    
1437    17. Compiling pcretest on Windows with readline support failed without the
1438        following two fixes: (1) Make the unistd.h include conditional on
1439        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1440    
1441    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1442        ncurses library to be included for pcretest when ReadLine support is
1443        requested, but also to allow for it to be overridden. This patch came from
1444        Daniel Bergström.
1445    
1446    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1447        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1448        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1449        spotting this.
1450    
1451    
1452    Version 7.6 28-Jan-08
1453    ---------------------
1454    
1455    1.  A character class containing a very large number of characters with
1456        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1457        overflow.
1458    
1459    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1460        HAVE_LONG_LONG is not defined.
1461    
1462    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1463        bring them up to date and include new features. This patch includes:
1464    
1465        - Fixed PH's badly added libz and libbz2 support.
1466        - Fixed a problem with static linking.
1467        - Added pcredemo. [But later removed - see 7 below.]
1468        - Fixed dftables problem and added an option.
1469        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1470            HAVE_LONG_LONG.
1471        - Added readline support for pcretest.
1472        - Added an listing of the option settings after cmake has run.
1473    
1474    4.  A user submitted a patch to Makefile that makes it easy to create
1475        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1476        Makefile.am that cause it to include this special target, without
1477        affecting anything else. Note that the same mingw target plus all
1478        the other distribution libraries and programs are now supported
1479        when configuring with CMake (see 6 below) instead of with
1480        Configure/Make.
1481    
1482    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1483        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1484        exported in the Windows port". It has not yet been confirmed that the patch
1485        solves the problem, but it does no harm.
1486    
1487    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1488        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1489        with CMake, and also correct the comment about stack recursion.
1490    
1491    7.  Remove the automatic building of pcredemo from the ./configure system and
1492        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1493        of a program that users should build themselves after PCRE is installed, so
1494        building it automatically is not really right. What is more, it gave
1495        trouble in some build environments.
1496    
1497    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1498    
1499    
1500    Version 7.5 10-Jan-08
1501    ---------------------
1502    
1503    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1504        values in parens when parsing an RE using the C++ wrapper."
1505    
1506    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1507        Characters greater than 255 were excluded from the class instead of being
1508        included.
1509    
1510    3.  The same bug as (2) above applied to negated POSIX classes such as
1511        [:^space:].
1512    
1513    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1514        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1515        I have changed it.
1516    
1517    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1518        first named subpattern) and a construct such as (?&a) would reference the
1519        first named subpattern whose name started with "a" (in other words, the
1520        length check was missing). Both these problems are fixed. "Subpattern name
1521        expected" is now given for (?&) (a zero-length name), and this patch also
1522        makes it give the same error for \k'' (previously it complained that that
1523        was a reference to a non-existent subpattern).
1524    
1525    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1526        this is right because (?- can be followed by option settings as well as by
1527        digits. I have, however, made the messages clearer.
1528    
1529    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1530        than the number used in the conditional) now cause a compile-time error.
1531        This is actually not compatible with Perl, which accepts such patterns, but
1532        treats the conditional as always being FALSE (as PCRE used to), but it
1533        seems to me that giving a diagnostic is better.
1534    
1535    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1536        and messages.
1537    
1538    9.  Fix two occurrences of "backslash" in comments that should have been
1539        "backspace".
1540    
1541    10. Remove two redundant lines of code that can never be obeyed (their function
1542        was moved elsewhere).
1543    
1544    11. The program that makes PCRE's Unicode character property table had a bug
1545        which caused it to generate incorrect table entries for sequences of
1546        characters that have the same character type, but are in different scripts.
1547        It amalgamated them into a single range, with the script of the first of
1548        them. In other words, some characters were in the wrong script. There were
1549        thirteen such cases, affecting characters in the following ranges:
1550    
1551          U+002b0 - U+002c1
1552          U+0060c - U+0060d
1553          U+0061e - U+00612
1554          U+0064b - U+0065e
1555          U+0074d - U+0076d
1556          U+01800 - U+01805
1557          U+01d00 - U+01d77
1558          U+01d9b - U+01dbf
1559          U+0200b - U+0200f
1560          U+030fc - U+030fe
1561          U+03260 - U+0327f
1562          U+0fb46 - U+0fbb1
1563          U+10450 - U+1049d
1564    
1565    12. The -o option (show only the matching part of a line) for pcregrep was not
1566        compatible with GNU grep in that, if there was more than one match in a
1567        line, it showed only the first of them. It now behaves in the same way as
1568        GNU grep.
1569    
1570    13. If the -o and -v options were combined for pcregrep, it printed a blank
1571        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1572        does the same. The return code can be used to tell if there were any
1573        non-matching lines.
1574    
1575    14. Added --file-offsets and --line-offsets to pcregrep.
1576    
1577    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1578        infinitely looping recursion. The bug was that positive lookaheads were not
1579        being skipped when checking for a possible empty match (negative lookaheads
1580        and both kinds of lookbehind were skipped).
1581    
1582    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1583        inclusion of <windows.h> to before rather than after the definition of
1584        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1585    
1586    17. Specifying a possessive quantifier with a specific limit for a Unicode
1587        character property caused pcre_compile() to compile bad code, which led at
1588        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1589        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1590        caused the error; without that there was no problem.
1591    
1592    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1593    
1594    19. Added --enable-pcretest-libreadline.
1595    
1596    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1597        RE::GlobalReplace(). As a result, the number of replacements returned was
1598        double what it should be. I removed one of the increments, but Craig sent a
1599        later patch that removed the other one (the right fix) and added unit tests
1600        that check the return values (which was not done before).
1601    
1602    21. Several CMake things:
1603    
1604        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1605            the names libpcre and libpcreposix, not just pcre and pcreposix.
1606    
1607        (2) The above change means that pcretest and pcregrep are now correctly
1608            linked with the newly-built libraries, not previously installed ones.
1609    
1610        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1611    
1612    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1613        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1614        UTF-8 newline character). The key issue is that the pattern starts .*;
1615        this means that the match must be either at the beginning, or after a
1616        newline. The bug was in the code for advancing after a failed match and
1617        checking that the new position followed a newline. It was not taking
1618        account of UTF-8 characters correctly.
1619    
1620    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1621        character classes. PCRE was not treating the sequence [:...:] as a
1622        character class unless the ... were all letters. Perl, however, seems to
1623        allow any characters between [: and :], though of course it rejects as
1624        unknown any "names" that contain non-letters, because all the known class
1625        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1626        for example, whereas PCRE did not - it did not recognize a POSIX character
1627        class. This seemed a bit dangerous, so the code has been changed to be
1628        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1629        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1630        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1631        Perl does, and where it didn't before.
1632    
1633    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1634        Windows environments %n is disabled by default.
1635    
1636    
1637    Version 7.4 21-Sep-07
1638    ---------------------
1639    
1640    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1641        means that a class such as [\s] counted as "explicit reference to CR or
1642        LF". That isn't really right - the whole point of the change was to try to
1643        help when there was an actual mention of one of the two characters. So now
1644        the change happens only if \r or \n (or a literal CR or LF) character is
1645        encountered.
1646    
1647    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1648        of both had grown to the point where there were only 3 bits left.
1649        Fortunately, there was spare space in the data structure, and so I have
1650        moved the internal flags into a new 16-bit field to free up more option
1651        bits.
1652    
1653    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1654        but did not set the internal JCHANGED flag - either of these is enough to
1655        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1656        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1657        start sets both bits.
1658    
1659    4.  Added options (at build time, compile time, exec time) to change \R from
1660        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1661    
1662    5.  doc/pcresyntax.html was missing from the distribution.
1663    
1664    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1665        compatibility, even though it is no longer used.
1666    
1667    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1668        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1669        windows.h file is present (where different names are used). [This was
1670        reversed later after testing - see 16 below.]
1671    
1672    8.  Changed all #include <config.h> to #include "config.h". There were also
1673        some further <pcre.h> cases that I changed to "pcre.h".
1674    
1675    9.  When pcregrep was used with the --colour option, it missed the line ending
1676        sequence off the lines that it output.
1677    
1678    10. It was pointed out to me that arrays of string pointers cause lots of
1679        relocations when a shared library is dynamically loaded. A technique of
1680        using a single long string with a table of offsets can drastically reduce
1681        these. I have refactored PCRE in four places to do this. The result is
1682        dramatic:
1683    
1684          Originally:                          290
1685          After changing UCP table:            187
1686          After changing error message table:   43
1687          After changing table of "verbs"       36
1688          After changing table of Posix names   22
1689    
1690        Thanks to the folks working on Gregex for glib for this insight.
1691    
1692    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1693        unicode-properties was also set.
1694    
1695    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1696    
1697    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1698        checked only for CRLF.
1699    
1700    14. Added casts to pcretest.c to avoid compiler warnings.
1701    
1702    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1703    
1704    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1705        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1706        entirely. This removes changes made in 7 above.
1707    
1708    17. The CMake files have been updated, and there is now more information about
1709        building with CMake in the NON-UNIX-USE document.
1710    
1711    
1712    Version 7.3 28-Aug-07
1713    ---------------------
1714    
1715     1. In the rejigging of the build system that eventually resulted in 7.1, the
1716        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1717        brackets there is not right, since it causes compilers to look for an
1718        installed pcre.h, not the version that is in the source that is being
1719        compiled (which of course may be different). I have changed it back to:
1720    
1721          #include "pcre.h"
1722    
1723        I have a vague recollection that the change was concerned with compiling in
1724        different directories, but in the new build system, that is taken care of
1725        by the VPATH setting the Makefile.
1726    
1727     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1728        when the subject happened to end in the byte 0x85 (e.g. if the last
1729        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1730        characters but of course it shouldn't be taken as a newline when it is part
1731        of another character. The bug was that, for an unlimited repeat of . in
1732        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1733        characters when looking for a newline.
1734    
1735     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1736    
1737     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1738        in debug output.
1739    
1740     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1741        long printing in the pcrecpp unittest when running under MinGW.
1742    
1743     6. ESC_K was left out of the EBCDIC table.
1744    
1745     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1746        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1747        limit also applies to "virtual nesting" when a pattern is recursive, and in
1748        this case 1000 isn't so big. I have been able to remove this limit at the
1749        expense of backing off one optimization in certain circumstances. Normally,
1750        when pcre_exec() would call its internal match() function recursively and
1751        immediately return the result unconditionally, it uses a "tail recursion"
1752        feature to save stack. However, when a subpattern that can match an empty
1753        string has an unlimited repetition quantifier, it no longer makes this
1754        optimization. That gives it a stack frame in which to save the data for
1755        checking that an empty string has been matched. Previously this was taken
1756        from the 1000-entry workspace that had been reserved. So now there is no
1757        explicit limit, but more stack is used.
1758    
1759     8. Applied Daniel's patches to solve problems with the import/export magic
1760        syntax that is required for Windows, and which was going wrong for the
1761        pcreposix and pcrecpp parts of the library. These were overlooked when this
1762        problem was solved for the main library.
1763    
1764     9. There were some crude static tests to avoid integer overflow when computing
1765        the size of patterns that contain repeated groups with explicit upper
1766        limits. As the maximum quantifier is 65535, the maximum group length was
1767        set at 30,000 so that the product of these two numbers did not overflow a
1768        32-bit integer. However, it turns out that people want to use groups that
1769        are longer than 30,000 bytes (though not repeat them that many times).
1770        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1771        made it possible to implement the integer overflow checks in a much more
1772        dynamic way, which I have now done. The artificial limitation on group
1773        length has been removed - we now have only the limit on the total length of
1774        the compiled pattern, which depends on the LINK_SIZE setting.
1775    
1776    10. Fixed a bug in the documentation for get/copy named substring when
1777        duplicate names are permitted. If none of the named substrings are set, the
1778        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1779        empty string.
1780    
1781    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1782        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1783        because the ] is interpreted as the first data character and the
1784        terminating ] is not found. PCRE has been made compatible with Perl in this
1785        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1786        cause memory overwriting.
1787    
1788    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1789        string has been matched (to stop an infinite loop). It was not recognizing
1790        a conditional subpattern that could match an empty string if that
1791        subpattern was within another subpattern. For example, it looped when
1792        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1793        condition was not nested. This bug has been fixed.
1794    
1795    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1796        past the start of the subject in the presence of bytes with the top bit
1797        set, for example "\x8aBCD".
1798    
1799    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1800        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1801    
1802    14. Optimized (?!) to (*FAIL).
1803    
1804    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1805        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1806        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1807        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1808        does: it's just the validity check that is more restrictive.
1809    
1810    16. Inserted checks for integer overflows during escape sequence (backslash)
1811        processing, and also fixed erroneous offset values for syntax errors during
1812        backslash processing.
1813    
1814    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1815        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1816    
1817    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1818        caused an overrun.
1819    
1820    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1821        something other than just ASCII characters) inside a group that had an
1822        unlimited repeat caused a loop at compile time (while checking to see
1823        whether the group could match an empty string).
1824    
1825    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1826        [\P{Any}] did so. (Error in the code for printing property names.)
1827    
1828    21. An orphan \E inside a character class could cause a crash.
1829    
1830    22. A repeated capturing bracket such as (A)? could cause a wild memory
1831        reference during compilation.
1832    
1833    23. There are several functions in pcre_compile() that scan along a compiled
1834        expression for various reasons (e.g. to see if it's fixed length for look
1835        behind). There were bugs in these functions when a repeated \p or \P was
1836        present in the pattern. These operators have additional parameters compared
1837        with \d, etc, and these were not being taken into account when moving along
1838        the compiled data. Specifically:
1839    
1840        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1841            length.
1842    
1843        (b) An item such as \pL+ within a repeated group could cause crashes or
1844            loops.
1845    
1846        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1847            "reference to non-existent subpattern" error.
1848    
1849        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1850    
1851    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1852        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1853    
1854    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1855    
1856    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1857        character were causing crashes (broken optimization).
1858    
1859    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1860        \p or \P) caused a compile-time loop.
1861    
1862    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1863        break. For example, the unstudied pattern [\r\n]A does not match the string
1864        "\r\nA" because change 7.0/46 below moves the current point on by two
1865        characters after failing to match at the start. However, the pattern \nA
1866        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1867        the same is true. There doesn't seem any very clean way out of this, but
1868        what I have chosen to do makes the common cases work: PCRE now takes note
1869        of whether there can be an explicit match for \r or \n anywhere in the
1870        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1871        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1872        pattern has explicit CR or LF references.
1873    
1874    29. Added (*CR) etc for changing newline setting at start of pattern.
1875    
1876    
1877    Version 7.2 19-Jun-07
1878    ---------------------
1879    
1880     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1881        which is apparently normally available under Windows.
1882    
1883     2. Re-jig the pcregrep tests with different newline settings in an attempt
1884        to make them independent of the local environment's newline setting.
1885    
1886     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1887    
1888     4. Some of the "internals" tests were previously cut out when the link size
1889        was not 2, because the output contained actual offsets. The recent new
1890        "Z" feature of pcretest means that these can be cut out, making the tests
1891        usable with all link sizes.
1892    
1893     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1894        stack recursion. This gives a massive performance boost under BSD, but just
1895        a small improvement under Linux. However, it saves one field in the frame
1896        in all cases.
1897    
1898     6. Added more features from the forthcoming Perl 5.10:
1899    
1900        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1901            recursion call. It refers to the nth most recently opened parentheses.
1902    
1903        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1904            to be opened parentheses.
1905    
1906        (c) Conditions that refer to capturing parentheses can be specified
1907            relatively, for example, (?(-2)... or (?(+3)...
1908    
1909        (d) \K resets the start of the current match so that everything before
1910            is not part of it.
1911    
1912        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1913    
1914        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1915            reference syntax.
1916    
1917        (g) (?| introduces a group in which the numbering of parentheses in each
1918            alternative starts with the same number.
1919    
1920        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1921    
1922     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1923        PCRE_INFO_JCHANGED.
1924    
1925     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1926        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1927        for detecting groups that can match an empty string.
1928    
1929     9. A pattern with a very large number of alternatives (more than several
1930        hundred) was running out of internal workspace during the pre-compile
1931        phase, where pcre_compile() figures out how much memory will be needed. A
1932        bit of new cunning has reduced the workspace needed for groups with
1933        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1934        workspace instead of running out of the 4096 that are available.
1935    
1936    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1937    
1938    11. Applied patch from Google to remove an optimization that didn't quite work.
1939        The report of the bug said:
1940    
1941          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1942          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1943          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1944    
1945    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1946        it matched the wrong number of bytes.
1947    
1948    
1949    Version 7.1 24-Apr-07
1950    ---------------------
1951    
1952     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1953        that is more "standard", making use of automake and other Autotools. There
1954        is some re-arrangement of the files and adjustment of comments consequent
1955        on this.
1956    
1957     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1958        for recursive directory scanning broke on some systems because the files
1959        are not scanned in any specific order and on different systems the order
1960        was different. A call to "sort" has been inserted into RunGrepTest for the
1961        approprate test as a short-term fix. In the longer term there may be an
1962        alternative.
1963    
1964     3. I had an email from Eric Raymond about problems translating some of PCRE's
1965        man pages to HTML (despite the fact that I distribute HTML pages, some
1966        people do their own conversions for various reasons). The problems
1967        concerned the use of low-level troff macros .br and .in. I have therefore
1968        removed all such uses from the man pages (some were redundant, some could
1969        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1970        HTML has been updated to handle .nf/.fi and to complain if it encounters
1971        .br or .in.
1972    
1973     4. Updated comments in configure.ac that get placed in config.h.in and also
1974        arranged for config.h to be included in the distribution, with the name
1975        config.h.generic, for the benefit of those who have to compile without
1976        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1977    
1978     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1979        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1980        makevp.bat for the current PCRE, using the additional files
1981        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1982    
1983     6. A Windows user reported a minor discrepancy with test 2, which turned out
1984        to be caused by a trailing space on an input line that had got lost in his
1985        copy. The trailing space was an accident, so I've just removed it.
1986    
1987     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1988        that is needed.
1989    
1990     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1991        as "const" (a) because they are and (b) because it helps the PHP
1992        maintainers who have recently made a script to detect big data structures
1993        in the php code that should be moved to the .rodata section. I remembered
1994        to update Builducptable as well, so it won't revert if ucptable.h is ever
1995        re-created.
1996    
1997     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1998        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1999        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2000        support is not required. This saves 1.5-2K of code, which is important in
2001        some applications.
2002    
2003        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2004        so as not to refer to the tables, even though these functions will never be
2005        called when UTF-8 support is disabled. Otherwise there are problems with a
2006        shared library.
2007    
2008    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2009    
2010        (a) It was defining its arguments as char * instead of void *.
2011    
2012        (b) It was assuming that all moves were upwards in memory; this was true
2013            a long time ago when I wrote it, but is no longer the case.
2014    
2015        The emulated memove() is provided for those environments that have neither
2016        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2017        is clearly not the case, as these two bugs were recently reported.
2018    
2019    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2020        and Detrail to create the HTML documentation, the .txt form of the man
2021        pages, and it removes trailing spaces from listed files. It also creates
2022        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2023        case, it wraps all the #defines with #ifndefs. This script should be run
2024        before "make dist".
2025    
2026    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2027        with Unicode property support.
2028    
2029        (a) For a maximizing quantifier, if the two different cases of the
2030            character were of different lengths in their UTF-8 codings (there are
2031            some cases like this - I found 11), and the matching function had to
2032            back up over a mixture of the two cases, it incorrectly assumed they
2033            were both the same length.
2034    
2035        (b) When PCRE was configured to use the heap rather than the stack for
2036            recursion during matching, it was not correctly preserving the data for
2037            the other case of a UTF-8 character when checking ahead for a match
2038            while processing a minimizing repeat. If the check also involved
2039            matching a wide character, but failed, corruption could cause an
2040            erroneous result when trying to check for a repeat of the original
2041            character.
2042    
2043    13. Some tidying changes to the testing mechanism:
2044    
2045        (a) The RunTest script now detects the internal link size and whether there
2046            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2047            values substituted by "configure". (The RunGrepTest script already did
2048            this for UTF-8.) The configure.ac script no longer substitutes the
2049            relevant variables.
2050    
2051        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2052            with length and offset values. This means that the output is different
2053            for different internal link sizes. Test 2 is skipped for link sizes
2054            other than 2 because of this, bypassing the problem. Unfortunately,
2055            there was also a test in test 3 (the locale tests) that used /B and
2056            failed for link sizes other than 2. Rather than cut the whole test out,
2057            I have added a new /Z option to pcretest that replaces the length and
2058            offset values with spaces. This is now used to make test 3 independent
2059            of link size. (Test 2 will be tidied up later.)
2060    
2061    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2062        segmentation fault instead of returning the appropriate error message.
2063    
2064    15. In multiline mode when the newline sequence was set to "any", the pattern
2065        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2066        This doesn't seem right; it now treats the CRLF combination as the line
2067        ending, and so does not match in that case. It's only a pattern such as ^$
2068        that would hit this one: something like ^ABC$ would have failed after \r
2069        and then tried again after \r\n.
2070    
2071    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2072        in an attempt to make files that differ only in their line terminators
2073        compare equal. This works on Linux.
2074    
2075    17. Under certain error circumstances pcregrep might try to free random memory
2076        as it exited. This is now fixed, thanks to valgrind.
2077    
2078    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2079        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2080        was because its rules for how to advance for /g after matching an empty
2081        string at the end of a line did not allow for this case. They now check for
2082        it specially.
2083    
2084    20. pcretest is supposed to handle patterns and data of any length, by
2085        extending its buffers when necessary. It was getting this wrong when the
2086        buffer for a data line had to be extended.
2087    
2088    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2089        CRLF as a newline sequence.
2090    
2091    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2092        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2093        I have nevertheless tidied it up.
2094    
2095    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2096    
2097    24. Added a man page for pcre-config.
2098    
2099    
2100    Version 7.0 19-Dec-06
2101    ---------------------
2102    
2103     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2104        moving to gcc 4.1.1.
2105    
2106     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2107        sys/time.h, which is documented as needed for this function. It doesn't
2108        seem to matter on Linux, but it showed up on some releases of OS X.
2109    
2110     3. It seems that there are systems where bytes whose values are greater than
2111        127 match isprint() in the "C" locale. The "C" locale should be the
2112        default when a C program starts up. In most systems, only ASCII printing
2113        characters match isprint(). This difference caused the output from pcretest
2114        to vary, making some of the tests fail. I have changed pcretest so that:
2115    
2116        (a) When it is outputting text in the compiled version of a pattern, bytes
2117            other than 32-126 are always shown as hex escapes.
2118    
2119        (b) When it is outputting text that is a matched part of a subject string,
2120            it does the same, unless a different locale has been set for the match
2121            (using the /L modifier). In this case, it uses isprint() to decide.
2122    
2123     4. Fixed a major bug that caused incorrect computation of the amount of memory
2124        required for a compiled pattern when options that changed within the
2125        pattern affected the logic of the preliminary scan that determines the
2126        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2127        that the computed length was too small. The symptoms of this bug were
2128        either the PCRE error "internal error: code overflow" from pcre_compile(),
2129        or a glibc crash with a message such as "pcretest: free(): invalid next
2130        size (fast)". Examples of patterns that provoked this bug (shown in
2131        pcretest format) are:
2132    
2133          /(?-x: )/x
2134          /(?x)(?-x: \s*#\s*)/
2135          /((?i)[\x{c0}])/8
2136          /(?i:[\x{c0}])/8
2137    
2138        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2139        is now done differently.
2140    
2141     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2142        wrapper classes; (b) implement a new function in the C++ scanner that is
2143        more efficient than the old way of doing things because it avoids levels of
2144        recursion in the regex matching; (c) add a paragraph to the documentation
2145        for the FullMatch() function.
2146    
2147     6. The escape sequence \n was being treated as whatever was defined as
2148        "newline". Not only was this contrary to the documentation, which states
2149        that \n is character 10 (hex 0A), but it also went horribly wrong when
2150        "newline" was defined as CRLF. This has been fixed.
2151    
2152     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2153        was being set to -1 for the "end of line" case (supposedly a value that no
2154        character can have). Though this value is never used (the check for end of
2155        line is "zero bytes in current character"), it caused compiler complaints.
2156        I've changed it to 0xffffffff.
2157    
2158     8. In pcre_version.c, the version string was being built by a sequence of
2159        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2160        string (as it is for production releases) called a macro with an empty
2161        argument. The C standard says the result of this is undefined. The gcc
2162        compiler treats it as an empty string (which was what was wanted) but it is
2163        reported that Visual C gives an error. The source has been hacked around to
2164        avoid this problem.
2165    
2166     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2167        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2168        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2169        of them did).
2170    
2171    10. Originally, pcretest opened its input and output without "b"; then I was
2172        told that "b" was needed in some environments, so it was added for release
2173        5.0 to both the input and output. (It makes no difference on Unix-like
2174        systems.) Later I was told that it is wrong for the input on Windows. I've
2175        now abstracted the modes into two macros, to make it easier to fiddle with
2176        them, and removed "b" from the input mode under Windows.
2177    
2178    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2179    
2180    12. Added -help and --help to pcretest as an official way of being reminded
2181        of the options.
2182    
2183    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2184        and pcrecpp.cc because they annoy compilers at high warning levels.
2185    
2186    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2187    
2188    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2189        scripts are not C programs :-) and which was not noticed because it works
2190        on Linux.
2191    
2192    16. pcretest is supposed to handle any length of pattern and data line (as one
2193        line or as a continued sequence of lines) by extending its input buffer if
2194        necessary. This feature was broken for very long pattern lines, leading to
2195        a string of junk being passed to pcre_compile() if the pattern was longer
2196        than about 50K.
2197    
2198    17. I have done a major re-factoring of the way pcre_compile() computes the
2199        amount of memory needed for a compiled pattern. Previously, there was code
2200        that made a preliminary scan of the pattern in order to do this. That was
2201        OK when PCRE was new, but as the facilities have expanded, it has become
2202        harder and harder to keep it in step with the real compile phase, and there
2203        have been a number of bugs (see for example, 4 above). I have now found a
2204        cunning way of running the real compile function in a "fake" mode that
2205        enables it to compute how much memory it would need, while actually only
2206        ever using a few hundred bytes of working memory and without too many
2207        tests of the mode. This should make future maintenance and development
2208        easier. A side effect of this work is that the limit of 200 on the nesting
2209        depth of parentheses has been removed (though this was never a serious
2210        limitation, I suspect). However, there is a downside: pcre_compile() now
2211        runs more slowly than before (30% or more, depending on the pattern). I
2212        hope this isn't a big issue. There is no effect on runtime performance.
2213    
2214    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2215        newline (only possible for the last line of a file) and it was a
2216        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2217    
2218    19. Added additional timing features to pcretest. (1) The -tm option now times
2219        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2220        separate command line item, by a number that specifies the number of
2221        repeats to use when timing. The default is 50000; this gives better
2222        precision, but takes uncomfortably long for very large patterns.
2223    
2224    20. Extended pcre_study() to be more clever in cases where a branch of a
2225        subpattern has no definite first character. For example, (a*|b*)[cd] would
2226        previously give no result from pcre_study(). Now it recognizes that the
2227        first character must be a, b, c, or d.
2228    
2229    21. There was an incorrect error "recursive call could loop indefinitely" if
2230        a subpattern (or the entire pattern) that was being tested for matching an
2231        empty string contained only one non-empty item after a nested subpattern.
2232        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2233        incorrectly, because the \d was being skipped in the check.
2234    
2235    22. The pcretest program now has a new pattern option /B and a command line
2236        option -b, which is equivalent to adding /B to every pattern. This causes
2237        it to show the compiled bytecode, without the additional information that
2238        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2239        is the same as /B/I).
2240    
2241    23. A new optimization is now able automatically to treat some sequences such
2242        as a*b as a*+b. More specifically, if something simple (such as a character
2243        or a simple class like \d) has an unlimited quantifier, and is followed by
2244        something that cannot possibly match the quantified thing, the quantifier
2245        is automatically "possessified".
2246    
2247    24. A recursive reference to a subpattern whose number was greater than 39
2248        went wrong under certain circumstances in UTF-8 mode. This bug could also
2249        have affected the operation of pcre_study().
2250    
2251    25. Realized that a little bit of performance could be had by replacing
2252        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2253    
2254    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2255    
2256    27. Possessive quantifiers such as a++ were previously implemented by turning
2257        them into atomic groups such as ($>a+). Now they have their own opcodes,
2258        which improves performance. This includes the automatically created ones
2259        from 23 above.
2260    
2261    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2262        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2263        the first matched character to be a colon. This applied both to named and
2264        numbered groups.
2265    
2266    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2267    
2268    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2269        building PCRE on Windows easier, so I have included it in the distribution.
2270    
2271    31. There is now a check in pcretest against a ridiculously large number being
2272        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2273        loop, the loop is abandoned.
2274    
2275    32. Forward references to subpatterns in conditions such as (?(2)...) where
2276        subpattern 2 is defined later cause pcre_compile() to search forwards in
2277        the pattern for the relevant set of parentheses. This search went wrong
2278        when there were unescaped parentheses in a character class, parentheses
2279        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2280    
2281    33. "Subroutine" calls and backreferences were previously restricted to
2282        referencing subpatterns earlier in the regex. This restriction has now
2283        been removed.
2284    
2285    34. Added a number of extra features that are going to be in Perl 5.10. On the
2286        whole, these are just syntactic alternatives for features that PCRE had
2287        previously implemented using the Python syntax or my own invention. The
2288        other formats are all retained for compatibility.
2289    
2290        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2291            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2292            also .NET compatible.
2293    
2294        (b) A recursion or subroutine call to a named group can now be defined as
2295            (?&name) as well as (?P>name).
2296    
2297        (c) A backreference to a named group can now be defined as \k<name> or
2298            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2299            5.10, are also .NET compatible.
2300    
2301        (d) A conditional reference to a named group can now use the syntax
2302            (?(<name>) or (?('name') as well as (?(name).
2303    
2304        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2305            groups (named and numbered) that are never evaluated inline, but can be
2306            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2307            is always false. There may be only one alternative in such a group.
2308    
2309        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2310            as the simple (?(R). The condition is true only if the most recent
2311            recursion is that of the given number or name. It does not search out
2312            through the entire recursion stack.
2313    
2314        (g) The escape \gN or \g{N} has been added, where N is a positive or
2315            negative number, specifying an absolute or relative reference.
2316    
2317    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2318        some "unreachable code" warnings.
2319    
2320    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2321        things, this adds five new scripts.
2322    
2323    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2324        There were also incompatibilities regarding the handling of \Q..\E inside
2325        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2326        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2327    
2328    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2329        matches an empty string, and forcibly breaks the loop. There were bugs in
2330        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2331        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2332        separate and independent bugs (that affected different cases) have been
2333        fixed.
2334    
2335    39. Refactored the code to abolish the use of different opcodes for small
2336        capturing bracket numbers. This is a tidy that I avoided doing when I
2337        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2338        The new approach is not only tidier, it makes it possible to reduce the
2339        memory needed to fix the previous bug (38).
2340    
2341    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2342        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2343        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2344        mode.
2345    
2346    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2347        report.
2348    
2349    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2350        copy construction and assignment in the C++ wrapper.
2351    
2352    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2353        couple of bugs that could have given wrong results in the "--newline=crlf"
2354        case.
2355    
2356    44. Added a number of casts and did some reorganization of signed/unsigned int
2357        variables following suggestions from Dair Grant. Also renamed the variable
2358        "this" as "item" because it is a C++ keyword.
2359    
2360    45. Arranged for dftables to add
2361    
2362          #include "pcre_internal.h"
2363    
2364        to pcre_chartables.c because without it, gcc 4.x may remove the array
2365        definition from the final binary if PCRE is built into a static library and
2366        dead code stripping is activated.
2367    
2368    46. For an unanchored pattern, if a match attempt fails at the start of a
2369        newline sequence, and the newline setting is CRLF or ANY, and the next two
2370        characters are CRLF, advance by two characters instead of one.
2371    
2372    
2373    Version 6.7 04-Jul-06
2374    ---------------------
2375    
2376     1. In order to handle tests when input lines are enormously long, pcretest has
2377        been re-factored so that it automatically extends its buffers when
2378        necessary. The code is crude, but this _is_ just a test program. The
2379        default size has been increased from 32K to 50K.
2380    
2381     2. The code in pcre_study() was using the value of the re argument before
2382        testing it for NULL. (Of course, in any sensible call of the function, it
2383        won't be NULL.)
2384    
2385     3. The memmove() emulation function in pcre_internal.h, which is used on
2386        systems that lack both memmove() and bcopy() - that is, hardly ever -
2387        was missing a "static" storage class specifier.
2388    
2389     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2390        containing an extended class (one that cannot be represented by a bitmap
2391        because it contains high-valued characters or Unicode property items, e.g.
2392        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2393        pattern, but PCRE should not loop if you do not (it no longer does).
2394        [Detail: two cases were found: (a) a repeated subpattern containing an
2395        extended class; (b) a recursive reference to a subpattern that followed a
2396        previous extended class. It wasn't skipping over the extended class
2397        correctly when UTF-8 mode was not set.]
2398    
2399     5. A negated single-character class was not being recognized as fixed-length
2400        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2401        compile error "lookbehind assertion is not fixed length".
2402    
2403     6. The RunPerlTest auxiliary script was showing an unexpected difference
2404        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2405        write a Perl script that can interpret lines of an input file either as
2406        byte characters or as UTF-8, which is what "perltest" was being required to
2407        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2408        can't do is switch easily at run time between having the "use utf8;" pragma
2409        or not. In the end, I fudged it by using the RunPerlTest script to insert
2410        "use utf8;" explicitly for the UTF-8 tests.
2411    
2412     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2413        the end of the subject string, contrary to the documentation and to what
2414        Perl does. This was true of both matching functions. Now it matches only at
2415        the start of the subject and immediately after *internal* newlines.
2416    
2417     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2418        a pointer to an int instead of a pointer to an unsigned long int. This
2419        caused problems on 64-bit systems.
2420    
2421     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2422        instance of the 'standard' template library not being so standard".
2423    
2424    10. There was no check on the number of named subpatterns nor the maximum
2425        length of a subpattern name. The product of these values is used to compute
2426        the size of the memory block for a compiled pattern. By supplying a very
2427        long subpattern name and a large number of named subpatterns, the size
2428        computation could be caused to overflow. This is now prevented by limiting
2429        the length of names to 32 characters, and the number of named subpatterns
2430        to 10,000.
2431    
2432    11. Subpatterns that are repeated with specific counts have to be replicated in
2433        the compiled pattern. The size of memory for this was computed from the
2434        length of the subpattern and the repeat count. The latter is limited to
2435        65535, but there was no limit on the former, meaning that integer overflow
2436        could in principle occur. The compiled length of a repeated subpattern is
2437        now limited to 30,000 bytes in order to prevent this.
2438    
2439    12. Added the optional facility to have named substrings with the same name.
2440    
2441    13. Added the ability to use a named substring as a condition, using the
2442        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2443        are numbers (not recommended). Forward references are permitted.
2444    
2445    14. Added forward references in named backreferences (if you see what I mean).
2446    
2447    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2448        pattern could run off the end of the subject. For example, the pattern
2449        "(?s)(.{1,5})"8 did this with the subject "ab".
2450    
2451    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2452        PCRE_CASELESS was set when matching characters that were quantified with ?
2453        or *.
2454    
2455    17. A character class other than a single negated character that had a minimum
2456        but no maximum quantifier - for example [ab]{6,} - was not handled
2457        correctly by pce_dfa_exec(). It would match only one character.
2458    
2459    18. A valid (though odd) pattern that looked like a POSIX character
2460        class but used an invalid character after [ (for example [[,abc,]]) caused
2461        pcre_compile() to give the error "Failed: internal error: code overflow" or
2462        in some cases to crash with a glibc free() error. This could even happen if
2463        the pattern terminated after [[ but there just happened to be a sequence of
2464        letters, a binary zero, and a closing ] in the memory that followed.
2465    
2466    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2467        over the years. Originally (before any Unicode support), just the bottom 8
2468        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2469        output from "man perlunicode" includes this:
2470    
2471          The regular expression compiler produces polymorphic opcodes.  That
2472          is, the pattern adapts to the data and automatically switches to
2473          the Unicode character scheme when presented with Unicode data--or
2474          instead uses a traditional byte scheme when presented with byte
2475          data.
2476    
2477        Sadly, a wide octal escape does not cause a switch, and in a string with
2478        no other multibyte characters, these octal escapes are treated as before.
2479        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2480        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2481        Unicode string.
2482    
2483        I have not perpetrated such confusion in PCRE. Up till now, it took just
2484        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2485        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2486        translate to the appropriate multibyte character.
2487    
2488    29. Applied some refactoring to reduce the number of warnings from Microsoft
2489        and Borland compilers. This has included removing the fudge introduced
2490        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2491        a warning about an unused variable.
2492    
2493    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2494        characters since release 4.0, because Perl (from release 5.004) does not.
2495        [Or at least, is documented not to: some releases seem to be in conflict
2496        with the documentation.] However, when a pattern was studied with
2497        pcre_study() and all its branches started with \s, PCRE still included VT
2498        as a possible starting character. Of course, this did no harm; it just
2499        caused an unnecessary match attempt.
2500    
2501    22. Removed a now-redundant internal flag bit that recorded the fact that case
2502        dependency changed within the pattern. This was once needed for "required
2503        byte" processing, but is no longer used. This recovers a now-scarce options
2504        bit. Also moved the least significant internal flag bit to the most-
2505        significant bit of the word, which was not previously used (hangover from
2506        the days when it was an int rather than a uint) to free up another bit for
2507        the future.
2508    
2509    23. Added support for CRLF line endings as well as CR and LF. As well as the
2510        default being selectable at build time, it can now be changed at runtime
2511        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2512        specify that it is scanning data with non-default line endings.
2513    
2514    24. Changed the definition of CXXLINK to make it agree with the definition of
2515        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2516    
2517    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2518        recursions. This makes a big different to stack usage for some patterns.
2519    
2520    26. If a subpattern containing a named recursion or subroutine reference such
2521        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2522        the space required for the compiled pattern went wrong and gave too small a
2523        value. Depending on the environment, this could lead to "Failed: internal
2524        error: code overflow at offset 49" or "glibc detected double free or
2525        corruption" errors.
2526    
2527    27. Applied patches from Google (a) to support the new newline modes and (b) to
2528        advance over multibyte UTF-8 characters in GlobalReplace.
2529    
2530    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2531        difference for some implementation of PCRE in some Windows version.
2532    
2533    29. Added some extra testing facilities to pcretest:
2534    
2535        \q<number>   in a data line sets the "match limit" value
2536        \Q<number>   in a data line sets the "match recursion limt" value
2537        -S <number>  sets the stack size, where <number> is in megabytes
2538    
2539        The -S option isn't available for Windows.
2540    
2541    
2542    Version 6.6 06-Feb-06
2543    ---------------------
2544    
2545     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2546        in pcreposix.h. I have copied the definition from pcre.h.
2547    
2548     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2549        because pcre.h is no longer a built file.
2550    
2551     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2552        not normally included in the compiled code.
2553    
2554    
2555    Version 6.5 01-Feb-06
2556    ---------------------
2557    
2558     1. When using the partial match feature with pcre_dfa_exec(), it was not
2559        anchoring the second and subsequent partial matches at the new starting
2560        point. This could lead to incorrect results. For example, with the pattern
2561        /1234/, partially matching against "123" and then "a4" gave a match.
2562    
2563     2. Changes to pcregrep:
2564    
2565        (a) All non-match returns from pcre_exec() were being treated as failures
2566            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2567            error message is output. Some extra information is given for the
2568            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2569            probably the only errors that are likely to be caused by users (by
2570            specifying a regex that has nested indefinite repeats, for instance).
2571            If there are more than 20 of these errors, pcregrep is abandoned.
2572    
2573        (b) A binary zero was treated as data while matching, but terminated the
2574            output line if it was written out. This has been fixed: binary zeroes
2575            are now no different to any other data bytes.
2576    
2577        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2578            used to set a locale for matching. The --locale=xxxx long option has
2579            been added (no short equivalent) to specify a locale explicitly on the
2580            pcregrep command, overriding the environment variables.
2581    
2582        (d) When -B was used with -n, some line numbers in the output were one less
2583            than they should have been.
2584    
2585        (e) Added the -o (--only-matching) option.
2586    
2587        (f) If -A or -C was used with -c (count only), some lines of context were
2588            accidentally printed for the final match.
2589    
2590        (g) Added the -H (--with-filename) option.
2591    
2592        (h) The combination of options -rh failed to suppress file names for files
2593            that were found from directory arguments.
2594    
2595        (i) Added the -D (--devices) and -d (--directories) options.
2596    
2597        (j) Added the -F (--fixed-strings) option.
2598    
2599        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2600    
2601        (l) Added the --colo(u)r option.
2602    
2603        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2604            is not present by default.
2605    
2606     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2607        items such as (?R) or (?1), when the recursion could match a number of
2608        alternatives. If it matched one of the alternatives, but subsequently,
2609        outside the recursion, there was a failure, the code tried to back up into
2610        the recursion. However, because of the way PCRE is implemented, this is not
2611        possible, and the result was an incorrect result from the match.
2612    
2613        In order to prevent this happening, the specification of recursion has
2614        been changed so that all such subpatterns are automatically treated as
2615        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2616    
2617     4. I had overlooked the fact that, in some locales, there are characters for
2618        which isalpha() is true but neither isupper() nor islower() are true. In
2619        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2620        and ordfeminine) are like this. This affected the treatment of \w and \W
2621        when they appeared in character classes, but not when they appeared outside
2622        a character class. The bit map for "word" characters is now created
2623        separately from the results of isalnum() instead of just taking it from the
2624        upper, lower, and digit maps. (Plus the underscore character, of course.)
2625    
2626     5. The above bug also affected the handling of POSIX character classes such as
2627        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2628        permanent tables. Instead, the bit maps for such a class were previously
2629        created as the appropriate unions of the upper, lower, and digit bitmaps.
2630        Now they are created by subtraction from the [[:word:]] class, which has
2631        its own bitmap.
2632    
2633     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2634        It is created by subtracting the vertical space characters (\x09, \x0a,
2635        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2636        subtraction was done in the overall bitmap for a character class, meaning
2637        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2638        be recognized. This bug has been fixed.
2639    
2640     7. Patches from the folks at Google:
2641    
2642          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2643          real life, but is still worth protecting against".
2644    
2645          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2646          regular expressions".
2647    
2648          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2649          have it.
2650    
2651          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2652          "configure" and the latter not, in order to fix a problem somebody had
2653          with compiling the Arg class on HP-UX.
2654    
2655          (e) Improve the error-handling of the C++ wrapper a little bit.
2656    
2657          (f) New tests for checking recursion limiting.
2658    
2659     8. The pcre_memmove() function, which is used only if the environment does not
2660        have a standard memmove() function (and is therefore rarely compiled),
2661        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2662        returning a result (though PCRE never actually uses the result).
2663    
2664     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2665        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2666        returned instead of calling malloc() with an overflowing number that would
2667        most likely cause subsequent chaos.
2668    
2669    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2670    
2671    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2672        with this option is matched, the nmatch and pmatch options of regexec() are
2673        ignored.
2674    
2675    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2676        provided in case anyone wants to the the POSIX interface with UTF-8
2677        strings.
2678    
2679    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2680        C++ linking (needed for some HP-UX environments).
2681    
2682    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2683        (unused parameter) and in the pcre_printint() function (omitted "default"
2684        switch label when the default is to do nothing).
2685    
2686    15. Added some code to make it possible, when PCRE is compiled as a C++
2687        library, to replace subject pointers for pcre_exec() with a smart pointer
2688        class, thus making it possible to process discontinuous strings.
2689    
2690    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2691        much the same function. They were added by different people who were trying
2692        to make PCRE easy to compile on non-Unix systems. It has been suggested
2693        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2694        for compiling on Windows systems. I have therefore replaced it with
2695        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2696        defaults to "extern" for C or "extern C" for C++, which works fine on
2697        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2698        SCOPE with something explicit in config.h. In addition:
2699    
2700        (a) pcreposix.h still had just "extern" instead of either of these macros;
2701            I have replaced it with PCRE_DATA_SCOPE.
2702    
2703        (b) Functions such as _pcre_xclass(), which are internal to the library,
2704            but external in the C sense, all had PCRE_EXPORT in their definitions.
2705            This is apparently wrong for the Windows case, so I have removed it.
2706            (It makes no difference on Unix-like systems.)
2707    
2708    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2709        of recursive calls to match(). This is different to MATCH_LIMIT because
2710        that limits the total number of calls to match(), not all of which increase
2711        the depth of recursion. Limiting the recursion depth limits the amount of
2712        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2713        when PCRE is compiled, and changed at run time. A patch from Google adds
2714        this functionality to the C++ interface.
2715    
2716    18. Changes to the handling of Unicode character properties:
2717    
2718        (a) Updated the table to Unicode 4.1.0.
2719    
2720        (b) Recognize characters that are not in the table as "Cn" (undefined).
2721    
2722        (c) I revised the way the table is implemented to a much improved format
2723            which includes recognition of ranges. It now supports the ranges that
2724            are defined in UnicodeData.txt, and it also amalgamates other
2725            characters into ranges. This has reduced the number of entries in the
2726            table from around 16,000 to around 3,000, thus reducing its size
2727            considerably. I realized I did not need to use a tree structure after
2728            all - a binary chop search is just as efficient. Having reduced the
2729            number of entries, I extended their size from 6 bytes to 8 bytes to
2730            allow for more data.
2731    
2732        (d) Added support for Unicode script names via properties such as \p{Han}.
2733    
2734    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2735        matching that character.
2736    
2737    20. When matching a repeated Unicode property with a minimum greater than zero,
2738        (for example \pL{2,}), PCRE could look past the end of the subject if it
2739        reached it while seeking the minimum number of characters. This could
2740        happen only if some of the characters were more than one byte long, because
2741        there is a check for at least the minimum number of bytes.
2742    
2743    21. Refactored the implementation of \p and \P so as to be more general, to
2744        allow for more different types of property in future. This has changed the
2745        compiled form incompatibly. Anybody with saved compiled patterns that use
2746        \p or \P will have to recompile them.
2747    
2748    22. Added "Any" and "L&" to the supported property types.
2749    
2750    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2751        but give a compile time error if the value is greater than 0xff.
2752    
2753    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2754        accidentally not being installed or uninstalled.
2755    
2756    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2757        made were to insert the current release number. This seemed silly, because
2758        it made things harder for people building PCRE on systems that don't run
2759        "configure". I have turned pcre.h into a distributed file, no longer built
2760        by "configure", with the version identification directly included. There is
2761        no longer a pcre.h.in file.
2762    
2763        However, this change necessitated a change to the pcre-config script as
2764        well. It is built from pcre-config.in, and one of the substitutions was the
2765        release number. I have updated configure.ac so that ./configure now finds
2766        the release number by grepping pcre.h.
2767    
2768    26. Added the ability to run the tests under valgrind.
2769    
2770    
2771    Version 6.4 05-Sep-05
2772    ---------------------
2773    
2774     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2775        "--" to be printed when multiple files were scanned, even when none of the
2776        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2777        consider it to be a bug, and have restored the previous behaviour.
2778    
2779     2. A couple of code tidies to get rid of compiler warnings.
2780    
2781     3. The pcretest program used to cheat by referring to symbols in the library
2782        whose names begin with _pcre_. These are internal symbols that are not
2783        really supposed to be visible externally, and in some environments it is
2784        possible to suppress them. The cheating is now confined to including
2785        certain files from the library's source, which is a bit cleaner.
2786    
2787     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2788        file's purpose clearer.
2789    
2790     5. Reorganized pcre_ucp_findchar().
2791    
2792    
2793    Version 6.3 15-Aug-05
2794    ---------------------
2795    
2796     1. The file libpcre.pc.in did not have general read permission in the tarball.
2797    
2798     2. There were some problems when building without C++ support:
2799    
2800        (a) If C++ support was not built, "make install" and "make test" still
2801            tried to test it.
2802    
2803        (b) There were problems when the value of CXX was explicitly set. Some
2804            changes have been made to try to fix these, and ...
2805    
2806        (c) --disable-cpp can now be used to explicitly disable C++ support.
2807    
2808        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2809            backslash in a target when C++ was disabled. This confuses some
2810            versions of "make", apparently. Using an intermediate variable solves
2811            this. (Same for CPP_LOBJ.)
2812    
2813     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2814        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2815        necessary on certain architectures.
2816    
2817     4. Added a setting of -export-symbols-regex to the link command to remove
2818        those symbols that are exported in the C sense, but actually are local
2819        within the library, and not documented. Their names all begin with
2820        "_pcre_". This is not a perfect job, because (a) we have to except some
2821        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2822        available (and never for static libraries). I have made a note to try to
2823        find a way round (a) in the future.
2824    
2825    
2826    Version 6.2 01-Aug-05
2827    ---------------------
2828    
2829     1. There was no test for integer overflow of quantifier values. A construction
2830        such as {1111111111111111} would give undefined results. What is worse, if
2831        a minimum quantifier for a parenthesized subpattern overflowed and became
2832        negative, the calculation of the memory size went wrong. This could have
2833        led to memory overwriting.
2834    
2835     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2836    
2837     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2838        operating environments where this matters.
2839    
2840     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2841        PCRE options from within the C++ wrapper.
2842    
2843     5. Named capturing subpatterns were not being correctly counted when a pattern
2844        was compiled. This caused two problems: (a) If there were more than 100
2845        such subpatterns, the calculation of the memory needed for the whole
2846        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2847        back references of the form \12, where the number was greater than 9, were
2848        not recognized as back references, even though there were sufficient
2849        previous subpatterns.
2850    
2851     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2852        versions of gcc, e.g. 2.95.4.
2853    
2854    
2855    Version 6.1 21-Jun-05
2856    ---------------------
2857    
2858     1. There was one reference to the variable "posix" in pcretest.c that was not
2859        surrounded by "#if !defined NOPOSIX".
2860    
2861     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2862        the cross-check on the old pcre_info() function, for the benefit of the
2863        cut-down version of PCRE that is currently imported into Exim.
2864    
2865     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2866        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2867        patterns that start (?i) but I don't think that matters. The use of (?i) is
2868        just an example; this all applies to the other options as well.
2869    
2870     4. Since libtool seems to echo the compile commands it is issuing, the output
2871        from "make" can be reduced a bit by putting "@" in front of each libtool
2872        compile command.
2873    
2874     5. Patch from the folks at Google for configure.in to be a bit more thorough
2875        in checking for a suitable C++ installation before trying to compile the
2876        C++ stuff. This should fix a reported problem when a compiler was present,
2877        but no suitable headers.
2878    
2879     6. The man pages all had just "PCRE" as their title. I have changed them to
2880        be the relevant file name. I have also arranged that these names are
2881        retained in the file doc/pcre.txt, which is a concatenation in text format
2882        of all the man pages except the little individual ones for each function.
2883    
2884     7. The NON-UNIX-USE file had not been updated for the different set of source
2885        files that come with release 6. I also added a few comments about the C++
2886        wrapper.
2887    
2888    
2889    Version 6.0 07-Jun-05
2890    ---------------------
2891    
2892     1. Some minor internal re-organization to help with my DFA experiments.
2893    
2894     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2895        didn't matter for the library itself when fully configured, but did matter
2896        when compiling without UCP support, or within Exim, where the ucp files are
2897        not imported.
2898    
2899     3. Refactoring of the library code to split up the various functions into
2900        different source modules. The addition of the new DFA matching code (see
2901        below) to a single monolithic source would have made it really too
2902        unwieldy, quite apart from causing all the code to be include in a
2903        statically linked application, when only some functions are used. This is
2904        relevant even without the DFA addition now that patterns can be compiled in
2905        one application and matched in another.
2906    
2907        The downside of splitting up is that there have to be some external
2908        functions and data tables that are used internally in different modules of
2909        the library but which are not part of the API. These have all had their
2910        names changed to start with "_pcre_" so that they are unlikely to clash
2911        with other external names.
2912    
2913     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2914        a different (DFA) algorithm. Although it is slower than the original
2915        function, it does have some advantages for certain types of matching
2916        problem.
2917    
2918     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2919        including restarting after a partial match.
2920    
2921     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2922        defined when compiling for Windows was sent to me. I have put it into the
2923        code, though I have no means of testing or verifying it.
2924    
2925     7. Added the pcre_refcount() auxiliary function.
2926    
2927     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2928        match before or at the first newline in the subject string. In pcretest,
2929        the /f option on a pattern can be used to set this.
2930    
2931     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2932        would behave wrongly. This has been present in PCRE since release 4.0.
2933    
2934    10. A number of changes to the pcregrep command:
2935    
2936        (a) Refactored how -x works; insert ^(...)$ instead of setting
2937            PCRE_ANCHORED and checking the length, in preparation for adding
2938            something similar for -w.
2939    
2940        (b) Added the -w (match as a word) option.
2941    
2942        (c) Refactored the way lines are read and buffered so as to have more
2943            than one at a time available.
2944    
2945        (d) Implemented a pcregrep test script.
2946    
2947        (e) Added the -M (multiline match) option. This allows patterns to match
2948            over several lines of the subject. The buffering ensures that at least
2949            8K, or the rest of the document (whichever is the shorter) is available
2950            for matching (and similarly the previous 8K for lookbehind assertions).
2951    
2952        (f) Changed the --help output so that it now says
2953    
2954              -w, --word-regex(p)
2955    
2956            instead of two lines, one with "regex" and the other with "regexp"
2957            because that confused at least one person since the short forms are the
2958            same. (This required a bit of code, as the output is generated
2959            automatically from a table. It wasn't just a text change.)
2960    
2961        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2962            option but starts with a hyphen. Could be a pattern or a path name
2963            starting with a hyphen, for instance.
2964    
2965        (h) "-" can be given as a file name to represent stdin.
2966    
2967        (i) When file names are being printed, "(standard input)" is used for
2968            the standard input, for compatibility with GNU grep. Previously
2969            "<stdin>" was used.
2970    
2971        (j) The option --label=xxx can be used to supply a name to be used for
2972            stdin when file names are being printed. There is no short form.
2973    
2974        (k) Re-factored the options decoding logic because we are going to add
2975            two more options that take data. Such options can now be given in four
2976            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2977    
2978        (l) Added the -A, -B, and -C options for requesting that lines of context
2979            around matches be printed.
2980    
2981        (m) Added the -L option to print the names of files that do not contain
2982            any matching lines, that is, the complement of -l.
2983    
2984        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2985            continue to scan other files.
2986    
2987        (o) The -s option was incorrectly implemented. For compatibility with other
2988            greps, it now suppresses the error message for a non-existent or non-
2989            accessible file (but not the return code). There is a new option called
2990            -q that suppresses the output of matching lines, which was what -s was
2991            previously doing.
2992    
2993        (p) Added --include and --exclude options to specify files for inclusion
2994            and exclusion when recursing.
2995    
2996    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2997        Hopefully, it now does.
2998    
2999    12. Missing cast in pcre_study().
3000    
3001    13. Added an "uninstall" target to the makefile.
3002    
3003    14. Replaced "extern" in the function prototypes in Makefile.in with
3004        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3005        world, but is set differently for Windows.
3006    
3007    15. Added a second compiling function called pcre_compile2(). The only
3008        difference is that it has an extra argument, which is a pointer to an
3009        integer error code. When there is a compile-time failure, this is set
3010        non-zero, in addition to the error test pointer being set to point to an
3011        error message. The new argument may be NULL if no error number is required
3012        (but then you may as well call pcre_compile(), which is now just a
3013        wrapper). This facility is provided because some applications need a
3014        numeric error indication, but it has also enabled me to tidy up the way
3015        compile-time errors are handled in the POSIX wrapper.
3016    
3017    16. Added VPATH=.libs to the makefile; this should help when building with one
3018        prefix path and installing with another. (Or so I'm told by someone who
3019        knows more about this stuff than I do.)
3020    
3021    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3022        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3023        match everything, including newlines. This is not POSIX-compatible, but
3024        somebody wanted the feature. From pcretest it can be activated by using
3025        both the P and the s flags.
3026    
3027    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3028    
3029    19. libpcre.pc was being incorrectly installed as executable.
3030    
3031    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3032        it now also looks for '\r' so that it will work unmodified on Windows.
3033    
3034    21. Added Google's contributed C++ wrapper to the distribution.
3035    
3036    22. Added some untidy missing memory free() calls in pcretest, to keep
3037        Electric Fence happy when testing.
3038    
3039    
3040    
3041    Version 5.0 13-Sep-04
3042    ---------------------
3043    
3044     1. Internal change: literal characters are no longer packed up into items
3045        containing multiple characters in a single byte-string. Each character
3046        is now matched using a separate opcode. However, there may be more than one
3047        byte in the character in UTF-8 mode.
3048    
3049     2. The pcre_callout_block structure has two new fields: pattern_position and
3050        next_item_length. These contain the offset in the pattern to the next match
3051        item, and its length, respectively.
3052    
3053     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3054        insertion of callouts before each pattern item. Added the /C option to
3055        pcretest to make use of this.
3056    
3057     4. On the advice of a Windows user, the lines
3058    
3059          #if defined(_WIN32) || defined(WIN32)
3060          _setmode( _fileno( stdout ), 0x8000 );
3061          #endif  /* defined(_WIN32) || defined(WIN32) */
3062    
3063        have been added to the source of pcretest. This apparently does useful
3064        magic in relation to line terminators.
3065    
3066     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3067        for the benefit of those environments where the "b" makes a difference.
3068    
3069     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3070        to know about it. I have put a hack into configure.in that adds in code
3071        to set GCC=yes if CC=icc. This seems to end up at a point in the
3072        generated configure script that is early enough to affect the setting of
3073        compiler options, which is what is needed, but I have no means of testing
3074        whether it really works. (The user who reported this had patched the
3075        generated configure script, which of course I cannot do.)
3076    
3077        LATER: After change 22 below (new libtool files), the configure script
3078        seems to know about icc (and also ecc). Therefore, I have commented out
3079        this hack in configure.in.
3080    
3081     7. Added support for pkg-config (2 patches were sent in).
3082    
3083     8. Negated POSIX character classes that used a combination of internal tables
3084        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3085        [[:^ascii]]. Typically, they would match almost any characters. The other
3086        POSIX classes were not broken in this way.
3087    
3088     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3089        to find the match, as PCRE was deluded into thinking that the match had to
3090        start at the start point or following a newline. The same bug applied to
3091        patterns with negative forward assertions or any backward assertions
3092        preceding ".*" at the start, unless the pattern required a fixed first
3093        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3094    
3095    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3096        starting at the last subject character, bytes beyond the end of the subject
3097        string were read.
3098    
3099    11. Renamed the variable "class" as "classbits" to make life easier for C++
3100        users. (Previously there was a macro definition, but it apparently wasn't
3101        enough.)
3102    
3103    12. Added the new field "tables" to the extra data so that tables can be passed
3104        in at exec time, or the internal tables can be re-selected. This allows
3105        a compiled regex to be saved and re-used at a later time by a different
3106        program that might have everything at different addresses.
3107    
3108    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3109        -R library as well as a -L library.
3110    
3111    14. The debugging options of pcretest (-d on the command line or D on a
3112        pattern) showed incorrect output for anything following an extended class
3113        that contained multibyte characters and which was followed by a quantifier.
3114    
3115    15. Added optional support for general category Unicode character properties
3116        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3117        support. It adds about 90K to the size of the library. The meanings of the
3118        inbuilt class escapes such as \d and \s have NOT been changed.
3119    
3120    16. Updated pcredemo.c to include calls to free() to release the memory for the
3121        compiled pattern.
3122    
3123    17. The generated file chartables.c was being created in the source directory
3124        instead of in the building directory. This caused the build to fail if the
3125        source directory was different from the building directory, and was
3126        read-only.
3127    
3128    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3129        file. No doubt somebody will tell me if they don't make sense... Also added
3130        Dan Mooney's comments about building on OpenVMS.
3131    
3132    19. Added support for partial matching via the PCRE_PARTIAL option for
3133        pcre_exec() and the \P data escape in pcretest.
3134    
3135    20. Extended pcretest with 3 new pattern features:
3136    
3137        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3138              write the compiled pattern to the file whose name is "rest-of-line".
3139              This is a straight binary dump of the data, with the saved pointer to
3140              the character tables forced to be NULL. The study data, if any, is
3141              written too. After writing, pcretest reads a new pattern.
3142    
3143        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3144              compiled pattern from the given file. There must not be any
3145              occurrences of "<" in the file name (pretty unlikely); if there are,
3146              pcretest will instead treat the initial "<" as a pattern delimiter.
3147              After reading in the pattern, pcretest goes on to read data lines as
3148              usual.
3149    
3150        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3151              and 16-bit fields in a compiled pattern, to simulate a pattern that
3152              was compiled on a host of opposite endianness.
3153    
3154    21. The pcre-exec() function can now cope with patterns that were compiled on
3155        hosts of opposite endianness, with this restriction:
3156    
3157          As for any compiled expression that is saved and used later, the tables
3158          pointer field cannot be preserved; the extra_data field in the arguments
3159          to pcre_exec() should be used to pass in a tables address if a value
3160          other than the default internal tables were used at compile time.
3161    
3162    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3163        now diagnosed as an error. Previously, most of the time, a negative number
3164        would have been treated as zero, but if in addition "ovector" was passed as
3165        NULL, a crash could occur.
3166    
3167    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3168        new versions from the libtool 1.5 distribution (the last one is a copy of
3169        a file called libtool.m4). This seems to have fixed the need to patch
3170        "configure" to support Darwin 1.3 (which I used to do). However, I still
3171        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3172        workstation).
3173    
3174    24. Changed the PCRE licence to be the more standard "BSD" licence.
3175    
3176    
3177    Version 4.5 01-Dec-03
3178    ---------------------
3179    
3180     1. There has been some re-arrangement of the code for the match() function so
3181        that it can be compiled in a version that does not call itself recursively.
3182        Instead, it keeps those local variables that need separate instances for
3183        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3184        needs to "recurse". Keeping track of where control must go is done by means
3185        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3186        hide most of the details from the main code, and operates only if
3187        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3188        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3189        operating.
3190    
3191        To make it easier for callers to provide specially tailored get/free
3192        functions for this usage, two new functions, pcre_stack_malloc, and
3193        pcre_stack_free, are used. They are always called in strict stacking order,
3194        and the size of block requested is always the same.
3195    
3196        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3197        PCRE has been compiled to use the stack or the heap for recursion. The
3198        -C option of pcretest uses this to show which version is compiled.
3199    
3200        A new data escape \S, is added to pcretest; it causes the amounts of store
3201        obtained and freed by both kinds of malloc/free at match time to be added
3202        to the output.
3203    
3204     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3205        what's available on my current Linux desktop machine.
3206    
3207     3. When matching a UTF-8 string, the test for a valid string at the start has
3208        been extended. If start_offset is not zero, PCRE now checks that it points
3209        to a byte that is the start of a UTF-8 character. If not, it returns
3210        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3211        this is necessary because there may be backward assertions in the pattern.
3212        When matching the same subject several times, it may save resources to use
3213        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3214    
3215     4. The code for checking the validity of UTF-8 strings has been tightened so
3216        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3217        containing "overlong sequences".
3218    
3219     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3220        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3221        should have been "&", but it just so happened that all the cases this let
3222        through by mistake were picked up later in the function.
3223    
3224     6. I had used a variable called "isblank" - this is a C99 function, causing
3225        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3226    
3227     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3228        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3229        so the version is shown; (c) stop "make test" echoing "./RunTest".
3230    
3231     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3232    
3233     9. The prototype for memmove() for systems that don't have it was using
3234        size_t, but the inclusion of the header that defines size_t was later. I've
3235        moved the #includes for the C headers earlier to avoid this.
3236    
3237    10. Added some adjustments to the code to make it easier to compiler on certain
3238        special systems:
3239    
3240          (a) Some "const" qualifiers were missing.
3241          (b) Added the macro EXPORT before all exported functions; by default this
3242              is defined to be empty.
3243          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3244              that it reads its output file name as an argument instead of writing
3245              to the standard output and assuming this can be redirected.
3246    
3247    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3248        class containing characters with values greater than 255, PCRE compilation
3249        went into a loop.
3250    
3251    12. A recursive reference to a subpattern that was within another subpattern
3252        that had a minimum quantifier of zero caused PCRE to crash. For example,
3253        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3254        recursion. If the recursively-called subpattern itself had a zero repeat,
3255        that was OK.
3256    
3257    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3258        buffer into which it was copied (for escape processing) was still set at
3259        1024, so long lines caused crashes.
3260    
3261    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3262        "internal error: code overflow...". This applied to any character class
3263        that was followed by a possessive quantifier.
3264    
3265    15. Modified the Makefile to add libpcre.la as a prerequisite for
3266        libpcreposix.la because I was told this is needed for a parallel build to
3267        work.
3268    
3269    16. If a pattern that contained .* following optional items at the start was
3270        studied, the wrong optimizing data was generated, leading to matching
3271        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3272        matching string must start with a or b or c. The correct conclusion for
3273        this pattern is that a match can start with any character.
3274    
3275    
3276    Version 4.4 13-Aug-03
3277    ---------------------
3278    
3279     1. In UTF-8 mode, a character class containing characters with values between
3280        127 and 255 was not handled correctly if the compiled pattern was studied.
3281        In fixing this, I have also improved the studying algorithm for such
3282        classes (slightly).
3283    
3284     2. Three internal functions had redundant arguments passed to them. Removal
3285        might give a very teeny performance improvement.
3286    
3287     3. Documentation bug: the value of the capture_top field in a callout is *one
3288        more than* the number of the hightest numbered captured substring.
3289    
3290     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3291        in incorrectly linking with a previously installed version. They now link
3292        explicitly with libpcre.la.
3293    
3294     5. configure.in no longer needs to recognize Cygwin specially.
3295    
3296     6. A problem in pcre.in for Windows platforms is fixed.
3297    
3298     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3299        pcretest, it used to include the size of the study block as part of its
3300        output. Unfortunately, the structure contains a field that has a different
3301        size on different hardware architectures. This meant that the tests that
3302        showed this size failed. As the block is currently always of a fixed size,
3303        this information isn't actually particularly useful in pcretest output, so
3304        I have just removed it.
3305    
3306     8. Three pre-processor statements accidentally did not start in column 1.
3307        Sadly, there are *still* compilers around that complain, even though
3308        standard C has not required this for well over a decade. Sigh.
3309    
3310     9. In pcretest, the code for checking callouts passed small integers in the
3311        callout_data field, which is a void * field. However, some picky compilers
3312        complained about the casts involved for this on 64-bit systems. Now
3313        pcretest passes the address of the small integer instead, which should get
3314        rid of the warnings.
3315    
3316    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3317        both compile and run time, and gives an error if an invalid UTF-8 sequence
3318        is found. There is a option for disabling this check in cases where the
3319        string is known to be correct and/or the maximum performance is wanted.
3320    
3321    11. In response to a bug report, I changed one line in Makefile.in from
3322    
3323            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3324        to
3325            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3326    
3327        to look similar to other lines, but I have no way of telling whether this
3328        is the right thing to do, as I do not use Windows. No doubt I'll get told
3329        if it's wrong...
3330    
3331    
3332    Version 4.3 21-May-03
3333    ---------------------
3334    
3335    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3336       Makefile.
3337    
3338    2. Some refactoring to improve the quality of the code:
3339    
3340       (i)   The utf8_table... variables are now declared "const".
3341    
3342       (ii)  The code for \cx, which used the "case flipping" table to upper case
3343             lower case letters, now just substracts 32. This is ASCII-specific,
3344             but the whole concept of \cx is ASCII-specific, so it seems
3345             reasonable.
3346    
3347       (iii) PCRE was using its character types table to recognize decimal and
3348             hexadecimal digits in the pattern. This is silly, because it handles
3349             only 0-9, a-f, and A-F, but the character types table is locale-
3350             specific, which means strange things might happen. A private
3351             table is now used for this - though it costs 256 bytes, a table is
3352             much faster than multiple explicit tests. Of course, the standard
3353             character types table is still used for matching digits in subject
3354             strings against \d.
3355    
3356       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3357             ending in _t are). So I've renamed it as ESC_tee.
3358    
3359    3. The first argument for regexec() in the POSIX wrapper should have been
3360       defined as "const".
3361    
3362    4. Changed pcretest to use malloc() for its buffers so that they can be
3363       Electric Fenced for debugging.
3364    
3365    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3366       to read one or more bytes before the start of the subject string. Often this
3367       had no effect on PCRE's behaviour, but in some circumstances it could
3368       provoke a segmentation fault.
3369    
3370    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3371       to try to read one or more bytes before the start of the subject string.
3372    
3373    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3374       UTF-8 support could misbehave in various ways if the subject string
3375       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3376       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3377       back over UTF-8 characters.)
3378    
3379    
3380    Version 4.2 14-Apr-03
3381    ---------------------
3382    
3383    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3384    
3385    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3386         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3387         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3388         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3389         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3390           and BUILD_EXEEXT
3391         Note: automatic setting of the BUILD variables is not yet working
3392         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3393           compile-time but not at link-time
3394         [LINK]: use for linking executables only
3395         make different versions for Windows and non-Windows
3396         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3397           libraries
3398         [LINK_FOR_BUILD]: new variable
3399         [OBJEXT]: use throughout
3400         [EXEEXT]: use throughout
3401         <winshared>: new target
3402         <wininstall>: new target
3403         <dftables.o>: use native compiler
3404         <dftables>: use native linker
3405         <install>: handle Windows platform correctly
3406         <clean>: ditto
3407         <check>: ditto
3408         copy DLL to top builddir before testing
3409    
3410       As part of these changes, -no-undefined was removed again. This was reported
3411       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3412       in any case.
3413    
3414    3. Some tidies to get rid of compiler warnings:
3415    
3416       . In the match_data structure, match_limit was an unsigned long int, whereas
3417         match_call_count was an int. I've made them both unsigned long ints.
3418    
3419       . In pcretest the fact that a const uschar * doesn't automatically cast to
3420         a void * provoked a warning.
3421    
3422       . Turning on some more compiler warnings threw up some "shadow" variables
3423         and a few more missing casts.
3424    
3425    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3426       option, a class that contained a single character with a value between 128
3427       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3428    
3429    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3430       option, a class that contained several characters, but with at least one
3431       whose value was between 128 and 255 caused PCRE to crash.
3432    
3433    
3434    Version 4.1 12-Mar-03
3435    ---------------------
3436    
3437    1. Compiling with gcc -pedantic found a couple of places where casts were
3438    needed, and a string in dftables.c that was longer than standard compilers are
3439    required to support.
3440    
3441    2. Compiling with Sun's compiler found a few more places where the code could
3442    be tidied up in order to avoid warnings.
3443    
3444    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3445    first of these names is deprecated in the latest Autoconf in favour of the name
3446    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3447    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3448    analogy I have changed it to CFLAGS_FOR_BUILD.
3449    
3450    4. Added -no-undefined to the linking command in the Makefile, because this is
3451    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3452    linking step for the pcreposix library.
3453    
3454    5. PCRE was failing to diagnose the case of two named groups with the same
3455    name.
3456    
3457    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3458    literal character that is needed in the subject for a match, and scans along to
3459    ensure that it is present before embarking on the full matching process. This
3460    saves time in cases of nested unlimited repeats that are never going to match.
3461    Problem: the scan can take a lot of time if the subject is very long (e.g.
3462    megabytes), thus penalizing straightforward matches. It is now done only if the
3463    amount of subject to be scanned is less than 1000 bytes.
3464    
3465    7. A lesser problem with the same optimization is that it was recording the
3466    first character of an anchored pattern as "needed", thus provoking a search
3467    right along the subject, even when the first match of the pattern was going to
3468    fail. The "needed" character is now not set for anchored patterns, unless it
3469    follows something in the pattern that is of non-fixed length. Thus, it still
3470    fulfils its original purpose of finding quick non-matches in cases of nested
3471    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3472    
3473    
3474    Version 4.0 17-Feb-03
3475    ---------------------
3476    
3477    1. If a comment in an extended regex that started immediately after a meta-item
3478    extended to the end of string, PCRE compiled incorrect data. This could lead to
3479    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3480    
3481    2. Moved to autoconf 2.53 and libtool 1.4.2.
3482    
3483    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3484    the special perltest8 script is no longer needed - all the tests can be run
3485    from a single perltest script.
3486    
3487    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3488    by \s. It has now been removed in PCRE. This means it isn't recognized as
3489    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3490    class [:space:] *does* include VT, thereby creating a mess.
3491    
3492    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3493    space and tab.
3494    
3495    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3496    its new features into the main test script, reducing the number of scripts.
3497    
3498    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3499    were backward compatible, and made the (?i) apply to the whole pattern, as if
3500    /i were given. Now it behaves more logically, and applies the option setting
3501    only to what follows. PCRE has been changed to follow suit. However, if it
3502    finds options settings right at the start of the pattern, it extracts them into
3503    the global options, as before. Thus, they show up in the info data.
3504    
3505    8. Added support for the \Q...\E escape sequence. Characters in between are
3506    treated as literals. This is slightly different from Perl in that $ and @ are
3507    also handled as literals inside the quotes. In Perl, they will cause variable
3508    interpolation. Note the following examples:
3509    
3510        Pattern            PCRE matches      Perl matches
3511    
3512        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3513        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3514        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3515    
3516    For compatibility with Perl, \Q...\E sequences are recognized inside character
3517    classes as well as outside them.
3518    
3519    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3520    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3521    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3522    signed/unsigned warnings.
3523    
3524    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3525    option for pcretest, so I've replaced it by a simple function that does just
3526    that job.
3527    
3528    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3529    "pcregrep -".
3530    
3531    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3532    Java package. This provides some syntactic sugar for simple cases of what my
3533    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3534    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3535    item, you can use this simplified notation. Note that only makes sense with
3536    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3537    greediness, whatever the setting of the PCRE_UNGREEDY option.
3538    
3539    13. A change of greediness default within a pattern was not taking effect at
3540    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3541    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3542    was abstracted outside.
3543    
3544    14. PCRE now supports the \G assertion. It is true when the current matching
3545    position is at the start point of the match. This differs from \A when the
3546    starting offset is non-zero. Used with the /g option of pcretest (or similar
3547    code), it works in the same way as it does for Perl's /g option. If all
3548    alternatives of a regex begin with \G, the expression is anchored to the start
3549    match position, and the "anchored" flag is set in the compiled expression.
3550    
3551    15. Some bugs concerning the handling of certain option changes within patterns
3552    have been fixed. These applied to options other than (?ims). For example,
3553    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3554    been the other way round. Some of this was related to change 7 above.
3555    
3556    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3557    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3558    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3559    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3560    
3561    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3562    mode. Unlike ".", it always matches newline, whatever the setting of
3563    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3564    assertions. Perl allows it, but it doesn't (in general) work because it can't
3565    calculate the length of the lookbehind. At least, that's the case for Perl
3566    5.8.0 - I've been told they are going to document that it doesn't work in
3567    future.
3568    
3569    18. Added an error diagnosis for escapes that PCRE does not support: these are
3570    \L, \l, \N, \P, \p, \U, \u, and \X.
3571    
3572    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3573    reading past the end of the pattern in cases such as /[abcd/.
3574    
3575    20. PCRE was getting more memory than necessary for patterns with classes that
3576    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3577    
3578    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3579    compiling PCRE for use with Virtual Pascal.
3580    
3581    22. Small fix to the Makefile to make it work properly if the build is done
3582    outside the source tree.
3583    
3584    23. Added a new extension: a condition to go with recursion. If a conditional
3585    subpattern starts with (?(R) the "true" branch is used if recursion has
3586    happened, whereas the "false" branch is used only at the top level.
3587    
3588    24. When there was a very long string of literal characters (over 255 bytes
3589    without UTF support, over 250 bytes with UTF support), the computation of how
3590    much memory was required could be incorrect, leading to segfaults or other
3591    strange effects.
3592    
3593    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3594    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3595    there was a subsequent back reference to those brackets. This meant that, for
3596    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3597    possible to check for precisely this case. All we can do is abandon the
3598    optimization if .* occurs inside capturing brackets when there are any back
3599    references whatsoever. (See below for a better fix that came later.)
3600    
3601    26. The handling of the optimization for finding the first character of a
3602    non-anchored pattern, and for finding a character that is required later in the
3603    match were failing in some cases. This didn't break the matching; it just
3604    failed to optimize when it could. The way this is done has been re-implemented.
3605    
3606    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3607    
3608    28. Added a new feature that provides some of the functionality that Perl
3609    provides with (?{...}). The facility is termed a "callout". The way it is done
3610    in PCRE is for the caller to provide an optional function, by setting
3611    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3612    global variable. By default it is unset, which disables all calling out. To get
3613    the function called, the regex must include (?C) at appropriate points. This
3614    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3615    This provides a means of identifying different callout points. When PCRE
3616    reaches such a point in the regex, if pcre_callout has been set, the external
3617    function is called. It is provided with data in a structure called
3618    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3619    matching continues; if it returns a non-zero value, the match at the current
3620    point fails. However, backtracking will occur if possible. [This was changed
3621    later and other features added - see item 49 below.]
3622    
3623    29. pcretest is upgraded to test the callout functionality. It provides a
3624    callout function that displays information. By default, it shows the start of
3625    the match and the current position in the text. There are some new data escapes
3626    to vary what happens:
3627    
3628        \C+         in addition, show current contents of captured substrings
3629        \C-         do not supply a callout function
3630        \C!n        return 1 when callout number n is reached
3631        \C!n!m      return 1 when callout number n is reached for the mth time
3632    
3633    30. If pcregrep was called with the -l option and just a single file name, it
3634    output "<stdin>" if a match was found, instead of the file name.
3635    
3636    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3637    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3638    pcre_exec(). This saves a malloc/free per call. The default value of
3639    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3640    when configuring.
3641    
3642    32. The default maximum size of a compiled pattern is 64K. There have been a
3643    few cases of people hitting this limit. The code now uses macros to handle the
3644    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3645    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3646    configuring. Tests 2 and 5 work only with 2-byte links because they output
3647    debugging information about compiled patterns.
3648    
3649    33. Internal code re-arrangements:
3650    
3651    (a) Moved the debugging function for printing out a compiled regex into
3652        its own source file (printint.c) and used #include to pull it into
3653        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3654        separate copies.
3655    
3656    (b) Defined the list of op-code names for debugging as a macro in
3657        internal.h so that it is next to the definition of the opcodes.
3658    
3659    (c) Defined a table of op-code lengths for simpler skipping along compiled
3660        code. This is again a macro in internal.h so that it is next to the
3661        definition of the opcodes.
3662    
3663    34. Added support for recursive calls to individual subpatterns, along the
3664    lines of Robin Houston's patch (but implemented somewhat differently).
3665    
3666    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3667    allow it to read and process whole directories in Win32. This code was
3668    contributed by Lionel Fourquaux; it has not been tested by me.
3669    
3670    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3671    used to name a group. Names consist of alphanumerics and underscores, and must
3672    be unique. Back references use the syntax (?P=name) and recursive calls use
3673    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3674    numbers. The function pcre_fullinfo() can be used after compilation to extract
3675    a name/number map. There are three relevant calls:
3676    
3677      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3678      PCRE_INFO_NAMECOUNT            yields the number of entries
3679      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3680    
3681    The map is a vector of fixed-size entries. The size of each entry depends on
3682    the length of the longest name used. The first two bytes of each entry are the
3683    group number, most significant byte first. There follows the corresponding
3684    name, zero terminated. The names are in alphabetical order.
3685    
3686    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3687    case instead of 255. Making it the same both with and without UTF-8 support
3688    means that the same test output works with both.
3689    
3690    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3691    calling malloc() with a zero argument.
3692    
3693    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3694    optimization. I've improved things by keeping a bitmap of backreferences with
3695    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3696    fact referenced, the optimization can be applied. It is unlikely that a
3697    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3698    the match to follow \n) will appear inside brackets with a number greater than
3699    31, but if it does, any back reference > 31 suppresses the optimization.
3700    
3701    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3702    of disabling numbered capturing parentheses. Any opening parenthesis that is
3703    not followed by ? behaves as if it were followed by ?: but named parentheses
3704    can still be used for capturing (and they will acquire numbers in the usual
3705    way).
3706    
3707    41. Redesigned the return codes from the match() function into yes/no/error so
3708    that errors can be passed back from deep inside the nested calls. A malloc
3709    failure while inside a recursive subpattern call now causes the
3710    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3711    
3712    42. It is now possible to set a limit on the number of times the match()
3713    function is called in a call to pcre_exec(). This facility makes it possible to
3714    limit the amount of recursion and backtracking, though not in a directly
3715    obvious way, because the match() function is used in a number of different
3716    circumstances. The count starts from zero for each position in the subject
3717    string (for non-anchored patterns). The default limit is, for compatibility, a
3718    large number, namely 10 000 000. You can change this in two ways:
3719    
3720    (a) When configuring PCRE before making, you can use --with-match-limit=n
3721        to set a default value for the compiled library.
3722    
3723    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3724        a different value is set. See 45 below.
3725    
3726    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3727    
3728    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3729    of things that can be changed at compile time. The first argument specifies
3730    what is wanted and the second points to where the information is to be placed.
3731    The current list of available information is:
3732    
3733      PCRE_CONFIG_UTF8
3734    
3735    The output is an integer that is set to one if UTF-8 support is available;
3736    otherwise it is set to zero.
3737    
3738      PCRE_CONFIG_NEWLINE
3739    
3740    The output is an integer that it set to the value of the code that is used for
3741    newline. It is either LF (10) or CR (13).
3742    
3743      PCRE_CONFIG_LINK_SIZE
3744    
3745    The output is an integer that contains the number of bytes used for internal
3746    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3747    
3748      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3749    
3750    The output is an integer that contains the threshold above which the POSIX
3751    interface uses malloc() for output vectors. See item 31 above.
3752    
3753      PCRE_CONFIG_MATCH_LIMIT
3754    
3755    The output is an unsigned integer that contains the default limit of the number
3756    of match() calls in a pcre_exec() execution. See 42 above.
3757    
3758    44. pcretest has been upgraded by the addition of the -C option. This causes it
3759    to extract all the available output from the new pcre_config() function, and to
3760    output it. The program then exits immediately.
3761    
3762    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3763    order to support additional features. One way would have been to define
3764    pcre_exec2() (for example) with extra arguments, but this would not have been
3765    extensible, and would also have required all calls to the original function to
3766    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3767    is used for passing in "extra" data from pcre_study().
3768    
3769    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3770    contains the following fields:
3771    
3772      flags         a bitmap indicating which of the following fields are set
3773      study_data    opaque data from pcre_study()
3774      match_limit   a way of specifying a limit on match() calls for a specific
3775                      call to pcre_exec()
3776      callout_data  data for callouts (see 49 below)
3777    
3778    The flag bits are also defined in pcre.h, and are
3779    
3780      PCRE_EXTRA_STUDY_DATA
3781      PCRE_EXTRA_MATCH_LIMIT
3782      PCRE_EXTRA_CALLOUT_DATA
3783    
3784    The pcre_study() function now returns one of these new pcre_extra blocks, with
3785    the actual study data pointed to by the study_data field, and the
3786    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3787    before. That is, this change is entirely upwards-compatible and requires no
3788    change to existing code.
3789    
3790    If you want to pass in additional data to pcre_exec(), you can either place it
3791    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3792    block.
3793    
3794    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3795    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3796    times with different match limits, until it finds the minimum value needed for
3797    pcre_exec() to complete. The value is then output. This can be instructive; for
3798    most simple matches the number is quite small, but for pathological cases it
3799    gets very large very quickly.
3800    
3801    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3802    returns the size of the data block pointed to by the study_data field in a
3803    pcre_extra block, that is, the value that was passed as the argument to
3804    pcre_malloc() when PCRE was getting memory in which to place the information
3805    created by pcre_study(). The fourth argument should point to a size_t variable.
3806    pcretest has been extended so that this information is shown after a successful
3807    pcre_study() call when information about the compiled regex is being displayed.
3808    
3809    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3810    because what follows is always an absolute path. (Later: it turns out that this