/[pcre]/code/branches/pcre16/ChangeLog
ViewVC logotype

Diff of /code/branches/pcre16/ChangeLog

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

revision 53 by nigel, Sat Feb 24 21:39:42 2007 UTC revision 749 by ph10, Fri Nov 18 10:36:45 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.21
5    ------------
6    
7    1.  Updating the JIT compiler.
8    
9    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10        are added as well.
11    
12    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14        calling _pcre_jit_exec. Some extra comments are added.
15    
16    4.  Mark settings inside atomic groups that do not contain any capturing
17        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18        was introduced by change 18 for 8.20.
19    
20    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21        ECMA-262 standard.
22    
23    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25        This bug was probably introduced by change 9 of 8.13.
26    
27    7.  While fixing 6 above, I noticed that a number of other items were being
28        incorrectly rejected as "not fixed length". This arose partly because newer
29        opcodes had not been added to the fixed-length checking code. I have (a)
30        corrected the bug and added tests for these items, and (b) arranged for an
31        error to occur if an unknown opcode is encountered while checking for fixed
32        length instead of just assuming "not fixed length". The items that were
33        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36    
37    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
38        being incorrectly compiled and would have given unpredicatble results.
39    
40    
41    Version 8.20 21-Oct-2011
42    ------------------------
43    
44    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
45        a POSIX class. After further experiments with Perl, which convinced me that
46        Perl has bugs and confusions, a closing square bracket is no longer allowed
47        in a POSIX name. This bug also affected patterns with classes that started
48        with full stops.
49    
50    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
51        captured substring, but while checking the failing first alternative,
52        substring 1 is temporarily captured. If the output vector supplied to
53        pcre_exec() was not big enough for this capture, the yield of the function
54        was still zero ("insufficient space for captured substrings"). This cannot
55        be totally fixed without adding another stack variable, which seems a lot
56        of expense for a edge case. However, I have improved the situation in cases
57        such as /(a)(b)x|abc/ matched against "abc", where the return code
58        indicates that fewer than the maximum number of slots in the ovector have
59        been set.
60    
61    3.  Related to (2) above: when there are more back references in a pattern than
62        slots in the output vector, pcre_exec() uses temporary memory during
63        matching, and copies in the captures as far as possible afterwards. It was
64        using the entire output vector, but this conflicts with the specification
65        that only 2/3 is used for passing back captured substrings. Now it uses
66        only the first 2/3, for compatibility. This is, of course, another edge
67        case.
68    
69    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
70        main code base, and can be used by building with --enable-jit. When this is
71        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
72        runtime --no-jit option is given.
73    
74    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
75        ovector, the return from the function was zero, implying that there were
76        other matches that did not fit. The correct "exactly full" value is now
77        returned.
78    
79    6.  If a subpattern that was called recursively or as a subroutine contained
80        (*PRUNE) or any other control that caused it to give a non-standard return,
81        invalid errors such as "Error -26 (nested recursion at the same subject
82        position)" or even infinite loops could occur.
83    
84    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
85        computing the minimum length on reaching *ACCEPT, and so ended up with the
86        wrong value of 1 rather than 0. Further investigation indicates that
87        computing a minimum subject length in the presence of *ACCEPT is difficult
88        (think back references, subroutine calls), and so I have changed the code
89        so that no minimum is registered for a pattern that contains *ACCEPT.
90    
91    8.  If (*THEN) was present in the first (true) branch of a conditional group,
92        it was not handled as intended. [But see 16 below.]
93    
94    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
95        Sheri Pierce.
96    
97    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
98        the first byte in a match must be "a".
99    
100    11. Change 17 for 8.13 increased the recursion depth for patterns like
101        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
102        pattern contains any instances of (*THEN). If it does not, the old
103        optimizations are restored. It would be nice to do this on a per-group
104        basis, but at the moment that is not feasible.
105    
106    12. In some environments, the output of pcretest -C is CRLF terminated. This
107        broke RunTest's code that checks for the link size. A single white space
108        character after the value is now allowed for.
109    
110    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
111        For "fr", it uses the Windows-specific input and output files.
112    
113    14. If (*THEN) appeared in a group that was called recursively or as a
114        subroutine, it did not work as intended. [But see next item.]
115    
116    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
117        pattern fragments (but not containing any | characters). If A and B are
118        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
119        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
120        D. In other words, Perl considers parentheses that do not contain any |
121        characters to be part of a surrounding alternative, whereas PCRE was
122        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
123        differently. PCRE now behaves in the same way as Perl, except in the case
124        of subroutine/recursion calls such as (?1) which have in any case always
125        been different (but PCRE had them first :-).
126    
127    16. Related to 15 above: Perl does not treat the | in a conditional group as
128        creating alternatives. Such a group is treated in the same way as an
129        ordinary group without any | characters when processing (*THEN). PCRE has
130        been changed to match Perl's behaviour.
131    
132    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
133        RunGrepTest script failed.
134    
135    18. Change 22 for version 13 caused atomic groups to use more stack. This is
136        inevitable for groups that contain captures, but it can lead to a lot of
137        stack use in large patterns. The old behaviour has been restored for atomic
138        groups that do not contain any capturing parentheses.
139    
140    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
141        suppress the check for a minimum subject length at run time. (If it was
142        given to pcre_exec() or pcre_dfa_exec() it did work.)
143    
144    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
145        fail to work when decoding hex characters in data strings in EBCDIC
146        environments.
147    
148    21. It appears that in at least one Mac OS environment, the isxdigit() function
149        is implemented as a macro that evaluates to its argument more than once,
150        contravening the C 90 Standard (I haven't checked a later standard). There
151        was an instance in pcretest which caused it to go wrong when processing
152        \x{...} escapes in subject strings. The has been rewritten to avoid using
153        things like p++ in the argument of isxdigit().
154    
155    
156    Version 8.13 16-Aug-2011
157    ------------------------
158    
159    1.  The Unicode data tables have been updated to Unicode 6.0.0.
160    
161    2.  Two minor typos in pcre_internal.h have been fixed.
162    
163    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
164        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
165        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
166    
167    4.  There were a number of related bugs in the code for matching backrefences
168        caselessly in UTF-8 mode when codes for the characters concerned were
169        different numbers of bytes. For example, U+023A and U+2C65 are an upper
170        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
171        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
172        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
173        2-byte code at the end of the subject (it thought there wasn't enough data
174        left).
175    
176    5.  Comprehensive information about what went wrong is now returned by
177        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
178        as the output vector has at least 2 elements. The offset of the start of
179        the failing character and a reason code are placed in the vector.
180    
181    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
182        now returned is for the first byte of the failing character, instead of the
183        last byte inspected. This is an incompatible change, but I hope it is small
184        enough not to be a problem. It makes the returned offset consistent with
185        pcre_exec() and pcre_dfa_exec().
186    
187    7.  pcretest now gives a text phrase as well as the error number when
188        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
189        failure, the offset and reason code are output.
190    
191    8.  When \R was used with a maximizing quantifier it failed to skip backwards
192        over a \r\n pair if the subsequent match failed. Instead, it just skipped
193        back over a single character (\n). This seems wrong (because it treated the
194        two characters as a single entity when going forwards), conflicts with the
195        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
196        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
197        has been changed.
198    
199    9.  Some internal refactoring has changed the processing so that the handling
200        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
201        time (the PCRE_DOTALL option was changed this way some time ago: version
202        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
203        which was always a bit of a fudge. It also means that there is one less
204        argument for the match() function, which reduces its stack requirements
205        slightly. This change also fixes an incompatibility with Perl: the pattern
206        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
207    
208    10. More internal refactoring has drastically reduced the number of recursive
209        calls to match() for possessively repeated groups such as (abc)++ when
210        using pcre_exec().
211    
212    11. While implementing 10, a number of bugs in the handling of groups were
213        discovered and fixed:
214    
215        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
216        (a|)*(?1) gave a compile-time internal error.
217        ((a|)+)+  did not notice that the outer group could match an empty string.
218        (^a|^)+   was not marked as anchored.
219        (.*a|.*)+ was not marked as matching at start or after a newline.
220    
221    12. Yet more internal refactoring has removed another argument from the match()
222        function. Special calls to this function are now indicated by setting a
223        value in a variable in the "match data" data block.
224    
225    13. Be more explicit in pcre_study() instead of relying on "default" for
226        opcodes that mean there is no starting character; this means that when new
227        ones are added and accidentally left out of pcre_study(), testing should
228        pick them up.
229    
230    14. The -s option of pcretest has been documented for ages as being an old
231        synonym of -m (show memory usage). I have changed it to mean "force study
232        for every regex", that is, assume /S for every regex. This is similar to -i
233        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
234        using it. It makes it easier to run collections of tests with and without
235        study enabled, and thereby test pcre_study() more easily. All the standard
236        tests are now run with and without -s (but some patterns can be marked as
237        "never study" - see 20 below).
238    
239    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
240        restoration of the capturing data to the outer values was not happening
241        correctly.
242    
243    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
244        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
245        pattern had matched an empty string, and so incorrectly returned a no
246        match.
247    
248    17. There was optimizing code for the last branch of non-capturing parentheses,
249        and also for the obeyed branch of a conditional subexpression, which used
250        tail recursion to cut down on stack usage. Unfortunately, now that there is
251        the possibility of (*THEN) occurring in these branches, tail recursion is
252        no longer possible because the return has to be checked for (*THEN). These
253        two optimizations have therefore been removed. [But see 8.20/11 above.]
254    
255    18. If a pattern containing \R was studied, it was assumed that \R always
256        matched two bytes, thus causing the minimum subject length to be
257        incorrectly computed because \R can also match just one byte.
258    
259    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
260        was incorrectly computed.
261    
262    20. If /S is present twice on a test pattern in pcretest input, it now
263        *disables* studying, thereby overriding the use of -s on the command line
264        (see 14 above). This is necessary for one or two tests to keep the output
265        identical in both cases.
266    
267    21. When (*ACCEPT) was used in an assertion that matched an empty string and
268        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
269    
270    22. When an atomic group that contained a capturing parenthesis was
271        successfully matched, but the branch in which it appeared failed, the
272        capturing was not being forgotten if a higher numbered group was later
273        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
274        group 1 to "a", when in fact it should be unset. This applied to multi-
275        branched capturing and non-capturing groups, repeated or not, and also to
276        positive assertions (capturing in negative assertions does not happen
277        in PCRE) and also to nested atomic groups.
278    
279    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
280        subject after a captured substring, to make it easier to tell which of a
281        number of identical substrings has been captured.
282    
283    24. The way atomic groups are processed by pcre_exec() has been changed so that
284        if they are repeated, backtracking one repetition now resets captured
285        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
286        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
287        "aaa". Previously, it would have been "a". As part of this code
288        refactoring, the way recursive calls are handled has also been changed.
289    
290    25. If an assertion condition captured any substrings, they were not passed
291        back unless some other capturing happened later. For example, if
292        (?(?=(a))a) was matched against "a", no capturing was returned.
293    
294    26. When studying a pattern that contained subroutine calls or assertions,
295        the code for finding the minimum length of a possible match was handling
296        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
297        group 1 called group 2 while simultaneously a separate group 2 called group
298        1). A stack overflow occurred in this case. I have fixed this by limiting
299        the recursion depth to 10.
300    
301    27. Updated RunTest.bat in the distribution to the version supplied by Tom
302        Fortmann. This supports explicit test numbers on the command line, and has
303        argument validation and error reporting.
304    
305    28. An instance of \X with an unlimited repeat could fail if at any point the
306        first character it looked at was a mark character.
307    
308    29. Some minor code refactoring concerning Unicode properties and scripts
309        should reduce the stack requirement of match() slightly.
310    
311    30. Added the '=' option to pcretest to check the setting of unused capturing
312        slots at the end of the pattern, which are documented as being -1, but are
313        not included in the return count.
314    
315    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
316        compiled something random. Now it gives a compile-time error (as does
317        Perl).
318    
319    32. A *MARK encountered during the processing of a positive assertion is now
320        recorded and passed back (compatible with Perl).
321    
322    33. If --only-matching or --colour was set on a pcregrep call whose pattern
323        had alternative anchored branches, the search for a second match in a line
324        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
325        matched the line "0102" twice. The same bug affected patterns that started
326        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
327        twice.
328    
329    34. Previously, PCRE did not allow quantification of assertions. However, Perl
330        does, and because of capturing effects, quantifying parenthesized
331        assertions may at times be useful. Quantifiers are now allowed for
332        parenthesized assertions.
333    
334    35. A minor code tidy in pcre_compile() when checking options for \R usage.
335    
336    36. \g was being checked for fancy things in a character class, when it should
337        just be a literal "g".
338    
339    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
340        appearance of a nested POSIX class supersedes an apparent external class.
341        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
342        unescaped square brackets may also appear as part of class names. For
343        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
344        more like Perl. (But see 8.20/1 above.)
345    
346    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
347        was because it thought it was \N{name}, which is not supported).
348    
349    39. Add minix to OS list not supporting the -S option in pcretest.
350    
351    40. PCRE tries to detect cases of infinite recursion at compile time, but it
352        cannot analyze patterns in sufficient detail to catch mutual recursions
353        such as ((?1))((?2)). There is now a runtime test that gives an error if a
354        subgroup is called recursively as a subpattern for a second time at the
355        same position in the subject string. In previous releases this might have
356        been caught by the recursion limit, or it might have run out of stack.
357    
358    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
359        happen only once. PCRE was, however incorrectly giving a compile time error
360        "recursive call could loop indefinitely" because it cannot analyze the
361        pattern in sufficient detail. The compile time test no longer happens when
362        PCRE is compiling a conditional subpattern, but actual runaway loops are
363        now caught at runtime (see 40 above).
364    
365    42. It seems that Perl allows any characters other than a closing parenthesis
366        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
367        has been changed to be the same.
368    
369    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
370        as not to get warnings when autogen.sh is called. Also changed
371        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
372    
373    44. To help people who use pcregrep to scan files containing exceedingly long
374        lines, the following changes have been made:
375    
376        (a) The default value of the buffer size parameter has been increased from
377            8K to 20K. (The actual buffer used is three times this size.)
378    
379        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
380            PCRE is built.
381    
382        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
383            to be set at run time.
384    
385        (d) Numerical values in pcregrep options can be followed by K or M, for
386            example --buffer-size=50K.
387    
388        (e) If a line being scanned overflows pcregrep's buffer, an error is now
389            given and the return code is set to 2.
390    
391    45. Add a pointer to the latest mark to the callout data block.
392    
393    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
394        partial match of an empty string instead of no match. This was specific to
395        the use of ".".
396    
397    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
398        complete match instead of a partial match. This bug was dependent on both
399        the PCRE_UTF8 and PCRE_DOTALL options being set.
400    
401    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
402        starting byte set, because \b was not being ignored.
403    
404    
405    Version 8.12 15-Jan-2011
406    ------------------------
407    
408    1.  Fixed some typos in the markup of the man pages, and wrote a script that
409        checks for such things as part of the documentation building process.
410    
411    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
412        --match-limit and --recursion-limit options (added for 8.11). In
413        particular, this made one of the standard tests fail. (The integer value
414        went into the wrong half of a long int.)
415    
416    3.  If the --colour option was given to pcregrep with -v (invert match), it
417        did strange things, either producing crazy output, or crashing. It should,
418        of course, ignore a request for colour when reporting lines that do not
419        match.
420    
421    4.  Another pcregrep bug caused similar problems if --colour was specified with
422        -M (multiline) and the pattern match finished with a line ending.
423    
424    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
425        matched in multiline mode, the following line was shown as part of the
426        match. This seems wrong, so I have changed it.
427    
428    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
429        the check for further matches in the same line (so they could be coloured)
430        to overrun the end of the current line. If another match was found, it was
431        incorrectly shown (and then shown again when found in the next line).
432    
433    7.  If pcregrep was compiled under Windows, there was a reference to the
434        function pcregrep_exit() before it was defined. I am assuming this was
435        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
436        reported by a user. I've moved the definition above the reference.
437    
438    
439    Version 8.11 10-Dec-2010
440    ------------------------
441    
442    1.  (*THEN) was not working properly if there were untried alternatives prior
443        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
444        backtracked to try for "b" instead of moving to the next alternative branch
445        at the same level (in this case, to look for "c"). The Perl documentation
446        is clear that when (*THEN) is backtracked onto, it goes to the "next
447        alternative in the innermost enclosing group".
448    
449    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
450        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
451        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
452        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
453        (*THEN).
454    
455    3.  If \s appeared in a character class, it removed the VT character from
456        the class, even if it had been included by some previous item, for example
457        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
458        of \s, but is part of the POSIX "space" class.)
459    
460    4.  A partial match never returns an empty string (because you can always
461        match an empty string at the end of the subject); however the checking for
462        an empty string was starting at the "start of match" point. This has been
463        changed to the "earliest inspected character" point, because the returned
464        data for a partial match starts at this character. This means that, for
465        example, /(?<=abc)def/ gives a partial match for the subject "abc"
466        (previously it gave "no match").
467    
468    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
469        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
470        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
471        has an implication that the given string is incomplete (because a partial
472        match is preferred over a full match). For this reason, these items now
473        give a partial match in this situation. [Aside: previously, the one case
474        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
475        match rather than a full match, which was wrong by the old rules, but is
476        now correct.]
477    
478    6.  There was a bug in the handling of #-introduced comments, recognized when
479        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
480        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
481        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
482        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
483        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
484        places in pcre_compile().
485    
486    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
487        comments when looking ahead for named forward references to subpatterns,
488        the only newline sequence it recognized was NL. It now handles newlines
489        according to the set newline convention.
490    
491    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
492        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
493        cater for a lack of strerror(). These oversights have been fixed.
494    
495    9.  Added --match-limit and --recursion-limit to pcregrep.
496    
497    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
498    
499    11. When the -o option was used, pcregrep was setting a return code of 1, even
500        when matches were found, and --line-buffered was not being honoured.
501    
502    12. Added an optional parentheses number to the -o and --only-matching options
503        of pcregrep.
504    
505    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
506        can match an empty string. The code to do it in pcretest and pcredemo
507        needed fixing:
508    
509        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
510            only one byte after an empty string match just before CRLF (this case
511            just got forgotten; "any" and "anycrlf" were OK).
512    
513        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
514            mode when an empty string match preceded an ASCII character followed by
515            a non-ASCII character. (The code for advancing by one character rather
516            than one byte was nonsense.)
517    
518        (c) The pcredemo.c sample program did not have any code at all to handle
519            the cases when CRLF is a valid newline sequence.
520    
521    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
522        as a starting offset was within the subject string. There is now a new
523        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
524        negative or greater than the length of the string. In order to test this,
525        pcretest is extended to allow the setting of negative starting offsets.
526    
527    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
528        starting offset points to the beginning of a UTF-8 character was
529        unnecessarily clumsy. I tidied it up.
530    
531    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
532        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
533    
534    17. Nobody had reported that the --include_dir option, which was added in
535        release 7.7 should have been called --include-dir (hyphen, not underscore)
536        for compatibility with GNU grep. I have changed it to --include-dir, but
537        left --include_dir as an undocumented synonym, and the same for
538        --exclude-dir, though that is not available in GNU grep, at least as of
539        release 2.5.4.
540    
541    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
542        characters from a string of bytes) have been redefined so as not to use
543        loops, in order to improve performance in some environments. At the same
544        time, I abstracted some of the common code into auxiliary macros to save
545        repetition (this should not affect the compiled code).
546    
547    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
548        compile-time error is now given if \c is not followed by an ASCII
549        character, that is, a byte less than 128. (In EBCDIC mode, the code is
550        different, and any byte value is allowed.)
551    
552    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
553        START_OPTIMIZE option, which is now allowed at compile time - but just
554        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
555        to pcregrep and other applications that have no direct access to PCRE
556        options. The new /Y option in pcretest sets this option when calling
557        pcre_compile().
558    
559    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
560        back references. Groups containing recursive back references were forced to
561        be atomic by that change, but in the case of named groups, the amount of
562        memory required was incorrectly computed, leading to "Failed: internal
563        error: code overflow". This has been fixed.
564    
565    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
566        pcretest.c, to avoid build problems in some Borland environments.
567    
568    
569    Version 8.10 25-Jun-2010
570    ------------------------
571    
572    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
573        THEN.
574    
575    2.  (*ACCEPT) was not working when inside an atomic group.
576    
577    3.  Inside a character class, \B is treated as a literal by default, but
578        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
579        causes the error). The code is unchanged, but I tidied the documentation.
580    
581    4.  Inside a character class, PCRE always treated \R and \X as literals,
582        whereas Perl faults them if its -w option is set. I have changed PCRE so
583        that it faults them when PCRE_EXTRA is set.
584    
585    5.  Added support for \N, which always matches any character other than
586        newline. (It is the same as "." when PCRE_DOTALL is not set.)
587    
588    6.  When compiling pcregrep with newer versions of gcc which may have
589        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
590        declared with attribute warn_unused_result" were given. Just casting the
591        result to (void) does not stop the warnings; a more elaborate fudge is
592        needed. I've used a macro to implement this.
593    
594    7.  Minor change to pcretest.c to avoid a compiler warning.
595    
596    8.  Added four artifical Unicode properties to help with an option to make
597        \s etc use properties (see next item). The new properties are: Xan
598        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
599    
600    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
601        use Unicode properties. (*UCP) at the start of a pattern can be used to set
602        this option. Modified pcretest to add /W to test this facility. Added
603        REG_UCP to make it available via the POSIX interface.
604    
605    10. Added --line-buffered to pcregrep.
606    
607    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
608        studied, and the match started with a letter with a code point greater than
609        127 whose first byte was different to the first byte of the other case of
610        the letter, the other case of this starting letter was not recognized
611        (#976).
612    
613    12. If a pattern that was studied started with a repeated Unicode property
614        test, for example, \p{Nd}+, there was the theoretical possibility of
615        setting up an incorrect bitmap of starting bytes, but fortunately it could
616        not have actually happened in practice until change 8 above was made (it
617        added property types that matched character-matching opcodes).
618    
619    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
620        possible starting bytes for non-anchored patterns.
621    
622    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
623        \R, and also a number of cases that involve Unicode properties, both
624        explicit and implicit when PCRE_UCP is set.
625    
626    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
627        input, it could crash or give wrong results if characters with values
628        greater than 0xc0 were present in the subject string. (Detail: it assumed
629        UTF-8 input when processing these items.)
630    
631    16. Added a lot of (int) casts to avoid compiler warnings in systems where
632        size_t is 64-bit (#991).
633    
634    17. Added a check for running out of memory when PCRE is compiled with
635        --disable-stack-for-recursion (#990).
636    
637    18. If the last data line in a file for pcretest does not have a newline on
638        the end, a newline was missing in the output.
639    
640    19. The default pcre_chartables.c file recognizes only ASCII characters (values
641        less than 128) in its various bitmaps. However, there is a facility for
642        generating tables according to the current locale when PCRE is compiled. It
643        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
644        characters, are recognized by isspace() and therefore were getting set in
645        these tables, and indeed these tables seem to approximate to ISO 8859. This
646        caused a problem in UTF-8 mode when pcre_study() was used to create a list
647        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
648        which of course cannot start UTF-8 characters. I have changed the code so
649        that only real ASCII characters (less than 128) and the correct starting
650        bytes for UTF-8 encodings are set for characters greater than 127 when in
651        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
652        altogether.)
653    
654    20. Added the /T option to pcretest so as to be able to run tests with non-
655        standard character tables, thus making it possible to include the tests
656        used for 19 above in the standard set of tests.
657    
658    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
659        reference to a subpattern the other side of a comment that contains an
660        opening parenthesis caused either an internal compiling error, or a
661        reference to the wrong subpattern.
662    
663    
664    Version 8.02 19-Mar-2010
665    ------------------------
666    
667    1.  The Unicode data tables have been updated to Unicode 5.2.0.
668    
669    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
670        configured.
671    
672    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
673        original author of that file, following a query about its status.
674    
675    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
676        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
677    
678    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
679        quantifier applied to a forward-referencing subroutine call, could compile
680        incorrect code or give the error "internal error: previously-checked
681        referenced subpattern not found".
682    
683    6.  Both MS Visual Studio and Symbian OS have problems with initializing
684        variables to point to external functions. For these systems, therefore,
685        pcre_malloc etc. are now initialized to local functions that call the
686        relevant global functions.
687    
688    7.  There were two entries missing in the vectors called coptable and poptable
689        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
690        I've fixed the data, and added a kludgy way of testing at compile time that
691        the lengths are correct (equal to the number of opcodes).
692    
693    8.  Following on from 7, I added a similar kludge to check the length of the
694        eint vector in pcreposix.c.
695    
696    9.  Error texts for pcre_compile() are held as one long string to avoid too
697        much relocation at load time. To find a text, the string is searched,
698        counting zeros. There was no check for running off the end of the string,
699        which could happen if a new error number was added without updating the
700        string.
701    
702    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
703    
704    11. \K was not working if it appeared in an atomic group or in a group that
705        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
706        \K is "not well defined" if used in an assertion. PCRE now accepts it if
707        the assertion is positive, but not if it is negative.
708    
709    12. Change 11 fortuitously reduced the size of the stack frame used in the
710        "match()" function of pcre_exec.c by one pointer. Forthcoming
711        implementation of support for (*MARK) will need an extra pointer on the
712        stack; I have reserved it now, so that the stack frame size does not
713        decrease.
714    
715    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
716        item in branch that calls a recursion is a subroutine call - as in the
717        second branch in the above example - was incorrectly given the compile-
718        time error "recursive call could loop indefinitely" because pcre_compile()
719        was not correctly checking the subroutine for matching a non-empty string.
720    
721    14. The checks for overrunning compiling workspace could trigger after an
722        overrun had occurred. This is a "should never occur" error, but it can be
723        triggered by pathological patterns such as hundreds of nested parentheses.
724        The checks now trigger 100 bytes before the end of the workspace.
725    
726    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
727    
728    
729    Version 8.01 19-Jan-2010
730    ------------------------
731    
732    1.  If a pattern contained a conditional subpattern with only one branch (in
733        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
734        computed the wrong minimum data length (which is of course zero for such
735        subpatterns). This could cause incorrect "no match" results.
736    
737    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
738        the pattern is reset in the first branch, pcre_compile() failed with
739        "internal error: code overflow at offset...". This happened only when
740        the reset was to the original external option setting. (An optimization
741        abstracts leading options settings into an external setting, which was the
742        cause of this.)
743    
744    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
745        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
746        assertion pattern did not match (meaning that the assertion was true), it
747        was incorrectly treated as false if the SKIP had been reached during the
748        matching. This also applied to assertions used as conditions.
749    
750    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
751        assertion subpattern, including such a pattern used as a condition,
752        unpredictable results occurred, instead of the error return
753        PCRE_ERROR_DFA_UITEM.
754    
755    5.  The C++ GlobalReplace function was not working like Perl for the special
756        situation when an empty string is matched. It now does the fancy magic
757        stuff that is necessary.
758    
759    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
760        removed. (These were left over from very, very early versions of PCRE.)
761    
762    7.  Some cosmetic changes to the code to make life easier when compiling it
763        as part of something else:
764    
765        (a) Change DEBUG to PCRE_DEBUG.
766    
767        (b) In pcre_compile(), rename the member of the "branch_chain" structure
768            called "current" as "current_branch", to prevent a collision with the
769            Linux macro when compiled as a kernel module.
770    
771        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
772            prevent a collision with the Linux macro when compiled as a kernel
773            module.
774    
775    8.  In pcre_compile() there are some checks for integer overflows that used to
776        cast potentially large values to (double). This has been changed to that
777        when building, a check for int64_t is made, and if it is found, it is used
778        instead, thus avoiding the use of floating point arithmetic. (There is no
779        other use of FP in PCRE.) If int64_t is not found, the fallback is to
780        double.
781    
782    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
783        2005 (difference between two addresses compared to an unsigned value).
784    
785    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
786        custom one, because of the following reported problem in Windows:
787    
788          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
789              under Win32.
790          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
791              therefore missing the function definition.
792          - The compiler thus generates a "C" signature for the test function.
793          - The linker fails to find the "C" function.
794          - PCRE fails to configure if asked to do so against libbz2.
795    
796    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
797        messages were output:
798    
799          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
800          rerunning libtoolize, to keep the correct libtool macros in-tree.
801          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
802    
803        I have done both of these things.
804    
805    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
806        most of the time, it *can* run out if it is given a pattern that contains a
807        runaway infinite recursion. I updated the discussion in the pcrestack man
808        page.
809    
810    13. Now that we have gone to the x.xx style of version numbers, the minor
811        version may start with zero. Using 08 or 09 is a bad idea because users
812        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
813        interpreted as invalid octal numbers. I've updated the previous comment in
814        configure.ac, and also added a check that gives an error if 08 or 09 are
815        used.
816    
817    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
818        causing partial matching to fail when the end of the subject matched \W
819        in a UTF-8 pattern where \W was quantified with a minimum of 3.
820    
821    15. There were some discrepancies between the declarations in pcre_internal.h
822        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
823        their definitions. The declarations used "const uschar *" and the
824        definitions used USPTR. Even though USPTR is normally defined as "const
825        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
826        reported that: "This difference in casting confuses some C++ compilers, for
827        example, SunCC recognizes above declarations as different functions and
828        generates broken code for hbpcre." I have changed the declarations to use
829        USPTR.
830    
831    16. GNU libtool is named differently on some systems. The autogen.sh script now
832        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
833        (FreeBSD).
834    
835    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
836        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
837        comment: "Figure out how to create a longlong from a string: strtoll and
838        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
839        instance, but it only takes 2 args instead of 3!"
840    
841    18. A subtle bug concerned with back references has been fixed by a change of
842        specification, with a corresponding code fix. A pattern such as
843        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
844        refers, was giving matches when it shouldn't. For example, xa=xaaa would
845        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
846        same bug. Such groups have to be quantified to be useful, or contained
847        inside another quantified group. (If there's no repetition, the reference
848        can never match.) The problem arises because, having left the group and
849        moved on to the rest of the pattern, a later failure that backtracks into
850        the group uses the captured value from the final iteration of the group
851        rather than the correct earlier one. I have fixed this in PCRE by forcing
852        any group that contains a reference to itself to be an atomic group; that
853        is, there cannot be any backtracking into it once it has completed. This is
854        similar to recursive and subroutine calls.
855    
856    
857    Version 8.00 19-Oct-09
858    ----------------------
859    
860    1.  The table for translating pcre_compile() error codes into POSIX error codes
861        was out-of-date, and there was no check on the pcre_compile() error code
862        being within the table. This could lead to an OK return being given in
863        error.
864    
865    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
866        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
867        in a Windows environment.
868    
869    3.  The pcregrep --count option prints the count for each file even when it is
870        zero, as does GNU grep. However, pcregrep was also printing all files when
871        --files-with-matches was added. Now, when both options are given, it prints
872        counts only for those files that have at least one match. (GNU grep just
873        prints the file name in this circumstance, but including the count seems
874        more useful - otherwise, why use --count?) Also ensured that the
875        combination -clh just lists non-zero counts, with no names.
876    
877    4.  The long form of the pcregrep -F option was incorrectly implemented as
878        --fixed_strings instead of --fixed-strings. This is an incompatible change,
879        but it seems right to fix it, and I didn't think it was worth preserving
880        the old behaviour.
881    
882    5.  The command line items --regex=pattern and --regexp=pattern were not
883        recognized by pcregrep, which required --regex pattern or --regexp pattern
884        (with a space rather than an '='). The man page documented the '=' forms,
885        which are compatible with GNU grep; these now work.
886    
887    6.  No libpcreposix.pc file was created for pkg-config; there was just
888        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
889    
890    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
891        when UCP support is not needed, by modifying the Python script that
892        generates it from Unicode data files. This should not matter if the module
893        is correctly used as a library, but I received one complaint about 50K of
894        unwanted data. My guess is that the person linked everything into his
895        program rather than using a library. Anyway, it does no harm.
896    
897    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
898        was a minimum greater than 1 for a wide character in a possessive
899        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
900        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
901        character. Chaos in the form of incorrect output or a compiling loop could
902        result.
903    
904    9.  The restrictions on what a pattern can contain when partial matching is
905        requested for pcre_exec() have been removed. All patterns can now be
906        partially matched by this function. In addition, if there are at least two
907        slots in the offset vector, the offset of the earliest inspected character
908        for the match and the offset of the end of the subject are set in them when
909        PCRE_ERROR_PARTIAL is returned.
910    
911    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
912        synonymous with PCRE_PARTIAL, for backwards compatibility, and
913        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
914        and may be more useful for multi-segment matching.
915    
916    11. Partial matching with pcre_exec() is now more intuitive. A partial match
917        used to be given if ever the end of the subject was reached; now it is
918        given only if matching could not proceed because another character was
919        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
920        string "Z", which now yields "no match" instead of "partial match". In the
921        case of pcre_dfa_exec(), "no match" is given if every matching path for the
922        final character ended with (*FAIL).
923    
924    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
925        if the pattern had a "must contain" character that was already found in the
926        earlier partial match, unless partial matching was again requested. For
927        example, with the pattern /dog.(body)?/, the "must contain" character is
928        "g". If the first part-match was for the string "dog", restarting with
929        "sbody" failed. This bug has been fixed.
930    
931    13. The string returned by pcre_dfa_exec() after a partial match has been
932        changed so that it starts at the first inspected character rather than the
933        first character of the match. This makes a difference only if the pattern
934        starts with a lookbehind assertion or \b or \B (\K is not supported by
935        pcre_dfa_exec()). It's an incompatible change, but it makes the two
936        matching functions compatible, and I think it's the right thing to do.
937    
938    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
939        so that the demonstration program is easily available in environments where
940        PCRE has not been installed from source.
941    
942    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
943        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
944        library.
945    
946    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
947        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
948        is not the first non-POSIX option to be added. Clearly some people find
949        these options useful.
950    
951    17. If a caller to the POSIX matching function regexec() passes a non-zero
952        value for nmatch with a NULL value for pmatch, the value of
953        nmatch is forced to zero.
954    
955    18. RunGrepTest did not have a test for the availability of the -u option of
956        the diff command, as RunTest does. It now checks in the same way as
957        RunTest, and also checks for the -b option.
958    
959    19. If an odd number of negated classes containing just a single character
960        interposed, within parentheses, between a forward reference to a named
961        subpattern and the definition of the subpattern, compilation crashed with
962        an internal error, complaining that it could not find the referenced
963        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
964        [The bug was that it was starting one character too far in when skipping
965        over the character class, thus treating the ] as data rather than
966        terminating the class. This meant it could skip too much.]
967    
968    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
969        /g option in pcretest when the pattern contains \K, which makes it possible
970        to have an empty string match not at the start, even when the pattern is
971        anchored. Updated pcretest and pcredemo to use this option.
972    
973    21. If the maximum number of capturing subpatterns in a recursion was greater
974        than the maximum at the outer level, the higher number was returned, but
975        with unset values at the outer level. The correct (outer level) value is
976        now given.
977    
978    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
979        PCRE did not set those parentheses (unlike Perl). I have now found a way to
980        make it do so. The string so far is captured, making this feature
981        compatible with Perl.
982    
983    23. The tests have been re-organized, adding tests 11 and 12, to make it
984        possible to check the Perl 5.10 features against Perl 5.10.
985    
986    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
987        pattern matches a fixed length string. PCRE did not allow this; now it
988        does. Neither allows recursion.
989    
990    25. I finally figured out how to implement a request to provide the minimum
991        length of subject string that was needed in order to match a given pattern.
992        (It was back references and recursion that I had previously got hung up
993        on.) This code has now been added to pcre_study(); it finds a lower bound
994        to the length of subject needed. It is not necessarily the greatest lower
995        bound, but using it to avoid searching strings that are too short does give
996        some useful speed-ups. The value is available to calling programs via
997        pcre_fullinfo().
998    
999    26. While implementing 25, I discovered to my embarrassment that pcretest had
1000        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1001        study optimizations had never been tested with that matching function.
1002        Oops. What is worse, even when it was passed study data, there was a bug in
1003        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1004        were also very few tests of studied patterns with pcre_dfa_exec().
1005    
1006    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1007        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1008        on the other side of the coin, they are no longer allowed to have different
1009        names, because these cannot be distinguished in PCRE, and this has caused
1010        confusion. (This is a difference from Perl.)
1011    
1012    28. When duplicate subpattern names are present (necessarily with different
1013        numbers, as required by 27 above), and a test is made by name in a
1014        conditional pattern, either for a subpattern having been matched, or for
1015        recursion in such a pattern, all the associated numbered subpatterns are
1016        tested, and the overall condition is true if the condition is true for any
1017        one of them. This is the way Perl works, and is also more like the way
1018        testing by number works.
1019    
1020    
1021    Version 7.9 11-Apr-09
1022    ---------------------
1023    
1024    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1025        (pcretest), all targets were linked against these libraries. This included
1026        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1027        libraries. This caused unwanted dependencies to be created. This problem
1028        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1029        pcretest is linked with readline.
1030    
1031    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1032        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1033        moved outside it again, because FALSE and TRUE are already defined in AIX,
1034        but BOOL is not.
1035    
1036    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1037        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1038    
1039    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1040        hyphens) following file names and line numbers when outputting matching
1041        lines. This is not true; no spaces are inserted. I have also clarified the
1042        wording for the --colour (or --color) option.
1043    
1044    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1045        was not coloured; this is different to GNU grep, so I have changed it to be
1046        the same.
1047    
1048    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1049        each matching line was coloured. Now it goes on to look for further matches
1050        of any of the test patterns, which is the same behaviour as GNU grep.
1051    
1052    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1053        doesn't make sense to accept an empty string match in pcregrep, so I have
1054        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1055        seems to be how GNU grep behaves.
1056    
1057    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1058        start or after a newline", because the conditional assertion was not being
1059        correctly handled. The rule now is that both the assertion and what follows
1060        in the first alternative must satisfy the test.
1061    
1062    9.  If auto-callout was enabled in a pattern with a conditional group whose
1063        condition was an assertion, PCRE could crash during matching, both with
1064        pcre_exec() and pcre_dfa_exec().
1065    
1066    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1067        used for matching.
1068    
1069    11. Unicode property support in character classes was not working for
1070        characters (bytes) greater than 127 when not in UTF-8 mode.
1071    
1072    12. Added the -M command line option to pcretest.
1073    
1074    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1075    
1076    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1077    
1078    16. Added comments and documentation about mis-use of no_arg in the C++
1079        wrapper.
1080    
1081    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1082        from Martin Jerabek that uses macro names for all relevant character and
1083        string constants.
1084    
1085    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1086        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1087        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1088        these, but not everybody uses configure.
1089    
1090    19. A conditional group that had only one branch was not being correctly
1091        recognized as an item that could match an empty string. This meant that an
1092        enclosing group might also not be so recognized, causing infinite looping
1093        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1094        with the subject "ab", where knowledge that the repeated group can match
1095        nothing is needed in order to break the loop.
1096    
1097    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1098        exec(), but without supplying a callout function, matching went wrong.
1099    
1100    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1101        leak if the size of the offset vector was greater than 30. When the vector
1102        is smaller, the saved offsets during recursion go onto a local stack
1103        vector, but for larger vectors malloc() is used. It was failing to free
1104        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1105        error, in fact).
1106    
1107    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1108        heapframe that is used only when UTF-8 support is enabled. This caused no
1109        problem, but was untidy.
1110    
1111    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1112        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1113        included within another project.
1114    
1115    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1116        slightly modified by me:
1117    
1118          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1119              not building pcregrep.
1120    
1121          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1122              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1123    
1124    25. Forward references, both numeric and by name, in patterns that made use of
1125        duplicate group numbers, could behave incorrectly or give incorrect errors,
1126        because when scanning forward to find the reference group, PCRE was not
1127        taking into account the duplicate group numbers. A pattern such as
1128        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1129    
1130    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1131        the feature of a custom pointer more persuasive (as requested by a user).
1132    
1133    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1134        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1135        pre-defined.
1136    
1137    28. Added support for (*UTF8) at the start of a pattern.
1138    
1139    29. Arrange for flags added by the "release type" setting in CMake to be shown
1140        in the configuration summary.
1141    
1142    
1143    Version 7.8 05-Sep-08
1144    ---------------------
1145    
1146    1.  Replaced UCP searching code with optimized version as implemented for Ad
1147        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1148        stage table and inline lookup instead of a function, giving speed ups of 2
1149        to 5 times on some simple patterns that I tested. Permission was given to
1150        distribute the MultiStage2.py script that generates the tables (it's not in
1151        the tarball, but is in the Subversion repository).
1152    
1153    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1154        scripts.
1155    
1156    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1157        a group with a zero qualifier. The result of the study could be incorrect,
1158        or the function might crash, depending on the pattern.
1159    
1160    4.  Caseless matching was not working for non-ASCII characters in back
1161        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1162        It now works when Unicode Property Support is available.
1163    
1164    5.  In pcretest, an escape such as \x{de} in the data was always generating
1165        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1166        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1167        truncation.
1168    
1169    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1170    
1171    7.  Added two (int) casts to pcregrep when printing the difference of two
1172        pointers, in case they are 64-bit values.
1173    
1174    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1175        test 2 if it fails.
1176    
1177    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1178        and a #define of that name to empty if it is not externally set. This is to
1179        allow users of MSVC to set it if necessary.
1180    
1181    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1182        the convenience functions in the pcre_get.c source file.
1183    
1184    11. An option change at the start of a pattern that had top-level alternatives
1185        could cause overwriting and/or a crash. This command provoked a crash in
1186        some environments:
1187    
1188          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1189    
1190        This potential security problem was recorded as CVE-2008-2371.
1191    
1192    12. For a pattern where the match had to start at the beginning or immediately
1193        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1194        pcre_dfa_exec() could read past the end of the passed subject if there was
1195        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1196        pcretest so that it places the subject at the end of its malloc-ed buffer.
1197    
1198    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1199        exec() might read past the end of the data buffer in UTF-8 mode.
1200    
1201    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1202        the data contained the byte 0x85 as part of a UTF-8 character within its
1203        first line. This applied both to normal and DFA matching.
1204    
1205    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1206        /^[^d]*?$/8 failed to match "abc".
1207    
1208    16. Added a missing copyright notice to pcrecpp_internal.h.
1209    
1210    17. Make it more clear in the documentation that values returned from
1211        pcre_exec() in ovector are byte offsets, not character counts.
1212    
1213    18. Tidied a few places to stop certain compilers from issuing warnings.
1214    
1215    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1216        supplied by Stefan Weber. I made a further small update for 7.8 because
1217        there is a change of source arrangements: the pcre_searchfuncs.c module is
1218        replaced by pcre_ucd.c.
1219    
1220    
1221    Version 7.7 07-May-08
1222    ---------------------
1223    
1224    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1225        a string to a long long, pretend we don't even have a long long." This is
1226        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1227    
1228    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1229        pre-7.6 versions, which defined a global no_arg variable instead of putting
1230        it in the RE class. (See also #8 below.)
1231    
1232    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1233        Lopes.
1234    
1235    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1236    
1237        (1) The include/exclude patterns were being applied to the whole pathnames
1238            of files, instead of just to the final components.
1239    
1240        (2) If there was more than one level of directory, the subdirectories were
1241            skipped unless they satisfied the include/exclude conditions. This is
1242            inconsistent with GNU grep (and could even be seen as contrary to the
1243            pcregrep specification - which I improved to make it absolutely clear).
1244            The action now is always to scan all levels of directory, and just
1245            apply the include/exclude patterns to regular files.
1246    
1247    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1248        --exclude_dir in the tests to avoid scanning .svn directories.
1249    
1250    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1251        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1252        doesn't support NULs in patterns.
1253    
1254    7.  Added some missing "const"s to declarations of static tables in
1255        pcre_compile.c and pcre_dfa_exec.c.
1256    
1257    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1258        caused by fix #2  above. (Subsequently also a second patch to fix the
1259        first patch. And a third patch - this was a messy problem.)
1260    
1261    9.  Applied Craig's patch to remove the use of push_back().
1262    
1263    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1264        matching function regexec().
1265    
1266    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1267        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1268        references. PCRE supports relative numbers with this syntax (I don't think
1269        Oniguruma does).
1270    
1271    12. Previously, a group with a zero repeat such as (...){0} was completely
1272        omitted from the compiled regex. However, this means that if the group
1273        was called as a subroutine from elsewhere in the pattern, things went wrong
1274        (an internal error was given). Such groups are now left in the compiled
1275        pattern, with a new opcode that causes them to be skipped at execution
1276        time.
1277    
1278    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1279        to the way PCRE behaves:
1280    
1281        (a) A lone ] character is dis-allowed (Perl treats it as data).
1282    
1283        (b) A back reference to an unmatched subpattern matches an empty string
1284            (Perl fails the current match path).
1285    
1286        (c) A data ] in a character class must be notated as \] because if the
1287            first data character in a class is ], it defines an empty class. (In
1288            Perl it is not possible to have an empty class.) The empty class []
1289            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1290            The negative empty class [^] matches any one character, independently
1291            of the DOTALL setting.
1292    
1293    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1294        non-existent subpattern following a character class starting with ']' and
1295        containing () gave an internal compiling error instead of "reference to
1296        non-existent subpattern". Fortunately, when the pattern did exist, the
1297        compiled code was correct. (When scanning forwards to check for the
1298        existencd of the subpattern, it was treating the data ']' as terminating
1299        the class, so got the count wrong. When actually compiling, the reference
1300        was subsequently set up correctly.)
1301    
1302    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1303        it was being rejected as not supported by pcre_dfa_exec(), even though
1304        other assertions are supported. I have made pcre_dfa_exec() support
1305        (*FAIL).
1306    
1307    16. The implementation of 13c above involved the invention of a new opcode,
1308        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1309        cannot be changed at match time, I realized I could make a small
1310        improvement to matching performance by compiling OP_ALLANY instead of
1311        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1312        on the OP_ANY path.
1313    
1314    17. Compiling pcretest on Windows with readline support failed without the
1315        following two fixes: (1) Make the unistd.h include conditional on
1316        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1317    
1318    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1319        ncurses library to be included for pcretest when ReadLine support is
1320        requested, but also to allow for it to be overridden. This patch came from
1321        Daniel Bergström.
1322    
1323    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1324        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1325        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1326        spotting this.
1327    
1328    
1329    Version 7.6 28-Jan-08
1330    ---------------------
1331    
1332    1.  A character class containing a very large number of characters with
1333        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1334        overflow.
1335    
1336    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1337        HAVE_LONG_LONG is not defined.
1338    
1339    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1340        bring them up to date and include new features. This patch includes:
1341    
1342        - Fixed PH's badly added libz and libbz2 support.
1343        - Fixed a problem with static linking.
1344        - Added pcredemo. [But later removed - see 7 below.]
1345        - Fixed dftables problem and added an option.
1346        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1347            HAVE_LONG_LONG.
1348        - Added readline support for pcretest.
1349        - Added an listing of the option settings after cmake has run.
1350    
1351    4.  A user submitted a patch to Makefile that makes it easy to create
1352        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1353        Makefile.am that cause it to include this special target, without
1354        affecting anything else. Note that the same mingw target plus all
1355        the other distribution libraries and programs are now supported
1356        when configuring with CMake (see 6 below) instead of with
1357        Configure/Make.
1358    
1359    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1360        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1361        exported in the Windows port". It has not yet been confirmed that the patch
1362        solves the problem, but it does no harm.
1363    
1364    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1365        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1366        with CMake, and also correct the comment about stack recursion.
1367    
1368    7.  Remove the automatic building of pcredemo from the ./configure system and
1369        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1370        of a program that users should build themselves after PCRE is installed, so
1371        building it automatically is not really right. What is more, it gave
1372        trouble in some build environments.
1373    
1374    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1375    
1376    
1377    Version 7.5 10-Jan-08
1378    ---------------------
1379    
1380    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1381        values in parens when parsing an RE using the C++ wrapper."
1382    
1383    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1384        Characters greater than 255 were excluded from the class instead of being
1385        included.
1386    
1387    3.  The same bug as (2) above applied to negated POSIX classes such as
1388        [:^space:].
1389    
1390    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1391        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1392        I have changed it.
1393    
1394    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1395        first named subpattern) and a construct such as (?&a) would reference the
1396        first named subpattern whose name started with "a" (in other words, the
1397        length check was missing). Both these problems are fixed. "Subpattern name
1398        expected" is now given for (?&) (a zero-length name), and this patch also
1399        makes it give the same error for \k'' (previously it complained that that
1400        was a reference to a non-existent subpattern).
1401    
1402    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1403        this is right because (?- can be followed by option settings as well as by
1404        digits. I have, however, made the messages clearer.
1405    
1406    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1407        than the number used in the conditional) now cause a compile-time error.
1408        This is actually not compatible with Perl, which accepts such patterns, but
1409        treats the conditional as always being FALSE (as PCRE used to), but it
1410        seems to me that giving a diagnostic is better.
1411    
1412    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1413        and messages.
1414    
1415    9.  Fix two occurrences of "backslash" in comments that should have been
1416        "backspace".
1417    
1418    10. Remove two redundant lines of code that can never be obeyed (their function
1419        was moved elsewhere).
1420    
1421    11. The program that makes PCRE's Unicode character property table had a bug
1422        which caused it to generate incorrect table entries for sequences of
1423        characters that have the same character type, but are in different scripts.
1424        It amalgamated them into a single range, with the script of the first of
1425        them. In other words, some characters were in the wrong script. There were
1426        thirteen such cases, affecting characters in the following ranges:
1427    
1428          U+002b0 - U+002c1
1429          U+0060c - U+0060d
1430          U+0061e - U+00612
1431          U+0064b - U+0065e
1432          U+0074d - U+0076d
1433          U+01800 - U+01805
1434          U+01d00 - U+01d77
1435          U+01d9b - U+01dbf
1436          U+0200b - U+0200f
1437          U+030fc - U+030fe
1438          U+03260 - U+0327f
1439          U+0fb46 - U+0fbb1
1440          U+10450 - U+1049d
1441    
1442    12. The -o option (show only the matching part of a line) for pcregrep was not
1443        compatible with GNU grep in that, if there was more than one match in a
1444        line, it showed only the first of them. It now behaves in the same way as
1445        GNU grep.
1446    
1447    13. If the -o and -v options were combined for pcregrep, it printed a blank
1448        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1449        does the same. The return code can be used to tell if there were any
1450        non-matching lines.
1451    
1452    14. Added --file-offsets and --line-offsets to pcregrep.
1453    
1454    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1455        infinitely looping recursion. The bug was that positive lookaheads were not
1456        being skipped when checking for a possible empty match (negative lookaheads
1457        and both kinds of lookbehind were skipped).
1458    
1459    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1460        inclusion of <windows.h> to before rather than after the definition of
1461        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1462    
1463    17. Specifying a possessive quantifier with a specific limit for a Unicode
1464        character property caused pcre_compile() to compile bad code, which led at
1465        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1466        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1467        caused the error; without that there was no problem.
1468    
1469    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1470    
1471    19. Added --enable-pcretest-libreadline.
1472    
1473    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1474        RE::GlobalReplace(). As a result, the number of replacements returned was
1475        double what it should be. I removed one of the increments, but Craig sent a
1476        later patch that removed the other one (the right fix) and added unit tests
1477        that check the return values (which was not done before).
1478    
1479    21. Several CMake things:
1480    
1481        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1482            the names libpcre and libpcreposix, not just pcre and pcreposix.
1483    
1484        (2) The above change means that pcretest and pcregrep are now correctly
1485            linked with the newly-built libraries, not previously installed ones.
1486    
1487        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1488    
1489    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1490        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1491        UTF-8 newline character). The key issue is that the pattern starts .*;
1492        this means that the match must be either at the beginning, or after a
1493        newline. The bug was in the code for advancing after a failed match and
1494        checking that the new position followed a newline. It was not taking
1495        account of UTF-8 characters correctly.
1496    
1497    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1498        character classes. PCRE was not treating the sequence [:...:] as a
1499        character class unless the ... were all letters. Perl, however, seems to
1500        allow any characters between [: and :], though of course it rejects as
1501        unknown any "names" that contain non-letters, because all the known class
1502        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1503        for example, whereas PCRE did not - it did not recognize a POSIX character
1504        class. This seemed a bit dangerous, so the code has been changed to be
1505        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1506        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1507        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1508        Perl does, and where it didn't before.
1509    
1510    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1511        Windows environments %n is disabled by default.
1512    
1513    
1514    Version 7.4 21-Sep-07
1515    ---------------------
1516    
1517    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1518        means that a class such as [\s] counted as "explicit reference to CR or
1519        LF". That isn't really right - the whole point of the change was to try to
1520        help when there was an actual mention of one of the two characters. So now
1521        the change happens only if \r or \n (or a literal CR or LF) character is
1522        encountered.
1523    
1524    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1525        of both had grown to the point where there were only 3 bits left.
1526        Fortunately, there was spare space in the data structure, and so I have
1527        moved the internal flags into a new 16-bit field to free up more option
1528        bits.
1529    
1530    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1531        but did not set the internal JCHANGED flag - either of these is enough to
1532        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1533        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1534        start sets both bits.
1535    
1536    4.  Added options (at build time, compile time, exec time) to change \R from
1537        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1538    
1539    5.  doc/pcresyntax.html was missing from the distribution.
1540    
1541    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1542        compatibility, even though it is no longer used.
1543    
1544    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1545        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1546        windows.h file is present (where different names are used). [This was
1547        reversed later after testing - see 16 below.]
1548    
1549    8.  Changed all #include <config.h> to #include "config.h". There were also
1550        some further <pcre.h> cases that I changed to "pcre.h".
1551    
1552    9.  When pcregrep was used with the --colour option, it missed the line ending
1553        sequence off the lines that it output.
1554    
1555    10. It was pointed out to me that arrays of string pointers cause lots of
1556        relocations when a shared library is dynamically loaded. A technique of
1557        using a single long string with a table of offsets can drastically reduce
1558        these. I have refactored PCRE in four places to do this. The result is
1559        dramatic:
1560    
1561          Originally:                          290
1562          After changing UCP table:            187
1563          After changing error message table:   43
1564          After changing table of "verbs"       36
1565          After changing table of Posix names   22
1566    
1567        Thanks to the folks working on Gregex for glib for this insight.
1568    
1569    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1570        unicode-properties was also set.
1571    
1572    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1573    
1574    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1575        checked only for CRLF.
1576    
1577    14. Added casts to pcretest.c to avoid compiler warnings.
1578    
1579    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1580    
1581    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1582        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1583        entirely. This removes changes made in 7 above.
1584    
1585    17. The CMake files have been updated, and there is now more information about
1586        building with CMake in the NON-UNIX-USE document.
1587    
1588    
1589    Version 7.3 28-Aug-07
1590    ---------------------
1591    
1592     1. In the rejigging of the build system that eventually resulted in 7.1, the
1593        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1594        brackets there is not right, since it causes compilers to look for an
1595        installed pcre.h, not the version that is in the source that is being
1596        compiled (which of course may be different). I have changed it back to:
1597    
1598          #include "pcre.h"
1599    
1600        I have a vague recollection that the change was concerned with compiling in
1601        different directories, but in the new build system, that is taken care of
1602        by the VPATH setting the Makefile.
1603    
1604     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1605        when the subject happened to end in the byte 0x85 (e.g. if the last
1606        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1607        characters but of course it shouldn't be taken as a newline when it is part
1608        of another character. The bug was that, for an unlimited repeat of . in
1609        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1610        characters when looking for a newline.
1611    
1612     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1613    
1614     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1615        in debug output.
1616    
1617     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1618        long printing in the pcrecpp unittest when running under MinGW.
1619    
1620     6. ESC_K was left out of the EBCDIC table.
1621    
1622     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1623        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1624        limit also applies to "virtual nesting" when a pattern is recursive, and in
1625        this case 1000 isn't so big. I have been able to remove this limit at the
1626        expense of backing off one optimization in certain circumstances. Normally,
1627        when pcre_exec() would call its internal match() function recursively and
1628        immediately return the result unconditionally, it uses a "tail recursion"
1629        feature to save stack. However, when a subpattern that can match an empty
1630        string has an unlimited repetition quantifier, it no longer makes this
1631        optimization. That gives it a stack frame in which to save the data for
1632        checking that an empty string has been matched. Previously this was taken
1633        from the 1000-entry workspace that had been reserved. So now there is no
1634        explicit limit, but more stack is used.
1635    
1636     8. Applied Daniel's patches to solve problems with the import/export magic
1637        syntax that is required for Windows, and which was going wrong for the
1638        pcreposix and pcrecpp parts of the library. These were overlooked when this
1639        problem was solved for the main library.
1640    
1641     9. There were some crude static tests to avoid integer overflow when computing
1642        the size of patterns that contain repeated groups with explicit upper
1643        limits. As the maximum quantifier is 65535, the maximum group length was
1644        set at 30,000 so that the product of these two numbers did not overflow a
1645        32-bit integer. However, it turns out that people want to use groups that
1646        are longer than 30,000 bytes (though not repeat them that many times).
1647        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1648        made it possible to implement the integer overflow checks in a much more
1649        dynamic way, which I have now done. The artificial limitation on group
1650        length has been removed - we now have only the limit on the total length of
1651        the compiled pattern, which depends on the LINK_SIZE setting.
1652    
1653    10. Fixed a bug in the documentation for get/copy named substring when
1654        duplicate names are permitted. If none of the named substrings are set, the
1655        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1656        empty string.
1657    
1658    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1659        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1660        because the ] is interpreted as the first data character and the
1661        terminating ] is not found. PCRE has been made compatible with Perl in this
1662        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1663        cause memory overwriting.
1664    
1665    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1666        string has been matched (to stop an infinite loop). It was not recognizing
1667        a conditional subpattern that could match an empty string if that
1668        subpattern was within another subpattern. For example, it looped when
1669        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1670        condition was not nested. This bug has been fixed.
1671    
1672    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1673        past the start of the subject in the presence of bytes with the top bit
1674        set, for example "\x8aBCD".
1675    
1676    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1677        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1678    
1679    14. Optimized (?!) to (*FAIL).
1680    
1681    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1682        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1683        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1684        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1685        does: it's just the validity check that is more restrictive.
1686    
1687    16. Inserted checks for integer overflows during escape sequence (backslash)
1688        processing, and also fixed erroneous offset values for syntax errors during
1689        backslash processing.
1690    
1691    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1692        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1693    
1694    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1695        caused an overrun.
1696    
1697    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1698        something other than just ASCII characters) inside a group that had an
1699        unlimited repeat caused a loop at compile time (while checking to see
1700        whether the group could match an empty string).
1701    
1702    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1703        [\P{Any}] did so. (Error in the code for printing property names.)
1704    
1705    21. An orphan \E inside a character class could cause a crash.
1706    
1707    22. A repeated capturing bracket such as (A)? could cause a wild memory
1708        reference during compilation.
1709    
1710    23. There are several functions in pcre_compile() that scan along a compiled
1711        expression for various reasons (e.g. to see if it's fixed length for look
1712        behind). There were bugs in these functions when a repeated \p or \P was
1713        present in the pattern. These operators have additional parameters compared
1714        with \d, etc, and these were not being taken into account when moving along
1715        the compiled data. Specifically:
1716    
1717        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1718            length.
1719    
1720        (b) An item such as \pL+ within a repeated group could cause crashes or
1721            loops.
1722    
1723        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1724            "reference to non-existent subpattern" error.
1725    
1726        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1727    
1728    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1729        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1730    
1731    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1732    
1733    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1734        character were causing crashes (broken optimization).
1735    
1736    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1737        \p or \P) caused a compile-time loop.
1738    
1739    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1740        break. For example, the unstudied pattern [\r\n]A does not match the string
1741        "\r\nA" because change 7.0/46 below moves the current point on by two
1742        characters after failing to match at the start. However, the pattern \nA
1743        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1744        the same is true. There doesn't seem any very clean way out of this, but
1745        what I have chosen to do makes the common cases work: PCRE now takes note
1746        of whether there can be an explicit match for \r or \n anywhere in the
1747        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1748        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1749        pattern has explicit CR or LF references.
1750    
1751    29. Added (*CR) etc for changing newline setting at start of pattern.
1752    
1753    
1754    Version 7.2 19-Jun-07
1755    ---------------------
1756    
1757     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1758        which is apparently normally available under Windows.
1759    
1760     2. Re-jig the pcregrep tests with different newline settings in an attempt
1761        to make them independent of the local environment's newline setting.
1762    
1763     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1764    
1765     4. Some of the "internals" tests were previously cut out when the link size
1766        was not 2, because the output contained actual offsets. The recent new
1767        "Z" feature of pcretest means that these can be cut out, making the tests
1768        usable with all link sizes.
1769    
1770     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1771        stack recursion. This gives a massive performance boost under BSD, but just
1772        a small improvement under Linux. However, it saves one field in the frame
1773        in all cases.
1774    
1775     6. Added more features from the forthcoming Perl 5.10:
1776    
1777        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1778            recursion call. It refers to the nth most recently opened parentheses.
1779    
1780        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1781            to be opened parentheses.
1782    
1783        (c) Conditions that refer to capturing parentheses can be specified
1784            relatively, for example, (?(-2)... or (?(+3)...
1785    
1786        (d) \K resets the start of the current match so that everything before
1787            is not part of it.
1788    
1789        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1790    
1791        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1792            reference syntax.
1793    
1794        (g) (?| introduces a group in which the numbering of parentheses in each
1795            alternative starts with the same number.
1796    
1797        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1798    
1799     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1800        PCRE_INFO_JCHANGED.
1801    
1802     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1803        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1804        for detecting groups that can match an empty string.
1805    
1806     9. A pattern with a very large number of alternatives (more than several
1807        hundred) was running out of internal workspace during the pre-compile
1808        phase, where pcre_compile() figures out how much memory will be needed. A
1809        bit of new cunning has reduced the workspace needed for groups with
1810        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1811        workspace instead of running out of the 4096 that are available.
1812    
1813    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1814    
1815    11. Applied patch from Google to remove an optimization that didn't quite work.
1816        The report of the bug said:
1817    
1818          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1819          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1820          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1821    
1822    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1823        it matched the wrong number of bytes.
1824    
1825    
1826    Version 7.1 24-Apr-07
1827    ---------------------
1828    
1829     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1830        that is more "standard", making use of automake and other Autotools. There
1831        is some re-arrangement of the files and adjustment of comments consequent
1832        on this.
1833    
1834     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1835        for recursive directory scanning broke on some systems because the files
1836        are not scanned in any specific order and on different systems the order
1837        was different. A call to "sort" has been inserted into RunGrepTest for the
1838        approprate test as a short-term fix. In the longer term there may be an
1839        alternative.
1840    
1841     3. I had an email from Eric Raymond about problems translating some of PCRE's
1842        man pages to HTML (despite the fact that I distribute HTML pages, some
1843        people do their own conversions for various reasons). The problems
1844        concerned the use of low-level troff macros .br and .in. I have therefore
1845        removed all such uses from the man pages (some were redundant, some could
1846        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1847        HTML has been updated to handle .nf/.fi and to complain if it encounters
1848        .br or .in.
1849    
1850     4. Updated comments in configure.ac that get placed in config.h.in and also
1851        arranged for config.h to be included in the distribution, with the name
1852        config.h.generic, for the benefit of those who have to compile without
1853        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1854    
1855     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1856        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1857        makevp.bat for the current PCRE, using the additional files
1858        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1859    
1860     6. A Windows user reported a minor discrepancy with test 2, which turned out
1861        to be caused by a trailing space on an input line that had got lost in his
1862        copy. The trailing space was an accident, so I've just removed it.
1863    
1864     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1865        that is needed.
1866    
1867     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1868        as "const" (a) because they are and (b) because it helps the PHP
1869        maintainers who have recently made a script to detect big data structures
1870        in the php code that should be moved to the .rodata section. I remembered
1871        to update Builducptable as well, so it won't revert if ucptable.h is ever
1872        re-created.
1873    
1874     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1875        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1876        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1877        support is not required. This saves 1.5-2K of code, which is important in
1878        some applications.
1879    
1880        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1881        so as not to refer to the tables, even though these functions will never be
1882        called when UTF-8 support is disabled. Otherwise there are problems with a
1883        shared library.
1884    
1885    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1886    
1887        (a) It was defining its arguments as char * instead of void *.
1888    
1889        (b) It was assuming that all moves were upwards in memory; this was true
1890            a long time ago when I wrote it, but is no longer the case.
1891    
1892        The emulated memove() is provided for those environments that have neither
1893        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1894        is clearly not the case, as these two bugs were recently reported.
1895    
1896    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1897        and Detrail to create the HTML documentation, the .txt form of the man
1898        pages, and it removes trailing spaces from listed files. It also creates
1899        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1900        case, it wraps all the #defines with #ifndefs. This script should be run
1901        before "make dist".
1902    
1903    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1904        with Unicode property support.
1905    
1906        (a) For a maximizing quantifier, if the two different cases of the
1907            character were of different lengths in their UTF-8 codings (there are
1908            some cases like this - I found 11), and the matching function had to
1909            back up over a mixture of the two cases, it incorrectly assumed they
1910            were both the same length.
1911    
1912        (b) When PCRE was configured to use the heap rather than the stack for
1913            recursion during matching, it was not correctly preserving the data for
1914            the other case of a UTF-8 character when checking ahead for a match
1915            while processing a minimizing repeat. If the check also involved
1916            matching a wide character, but failed, corruption could cause an
1917            erroneous result when trying to check for a repeat of the original
1918            character.
1919    
1920    13. Some tidying changes to the testing mechanism:
1921    
1922        (a) The RunTest script now detects the internal link size and whether there
1923            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1924            values substituted by "configure". (The RunGrepTest script already did
1925            this for UTF-8.) The configure.ac script no longer substitutes the
1926            relevant variables.
1927    
1928        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1929            with length and offset values. This means that the output is different
1930            for different internal link sizes. Test 2 is skipped for link sizes
1931            other than 2 because of this, bypassing the problem. Unfortunately,
1932            there was also a test in test 3 (the locale tests) that used /B and
1933            failed for link sizes other than 2. Rather than cut the whole test out,
1934            I have added a new /Z option to pcretest that replaces the length and
1935            offset values with spaces. This is now used to make test 3 independent
1936            of link size. (Test 2 will be tidied up later.)
1937    
1938    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1939        segmentation fault instead of returning the appropriate error message.
1940    
1941    15. In multiline mode when the newline sequence was set to "any", the pattern
1942        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1943        This doesn't seem right; it now treats the CRLF combination as the line
1944        ending, and so does not match in that case. It's only a pattern such as ^$
1945        that would hit this one: something like ^ABC$ would have failed after \r
1946        and then tried again after \r\n.
1947    
1948    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1949        in an attempt to make files that differ only in their line terminators
1950        compare equal. This works on Linux.
1951    
1952    17. Under certain error circumstances pcregrep might try to free random memory
1953        as it exited. This is now fixed, thanks to valgrind.
1954    
1955    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1956        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1957        was because its rules for how to advance for /g after matching an empty
1958        string at the end of a line did not allow for this case. They now check for
1959        it specially.
1960    
1961    20. pcretest is supposed to handle patterns and data of any length, by
1962        extending its buffers when necessary. It was getting this wrong when the
1963        buffer for a data line had to be extended.
1964    
1965    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1966        CRLF as a newline sequence.
1967    
1968    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1969        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1970        I have nevertheless tidied it up.
1971    
1972    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1973    
1974    24. Added a man page for pcre-config.
1975    
1976    
1977    Version 7.0 19-Dec-06
1978    ---------------------
1979    
1980     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1981        moving to gcc 4.1.1.
1982    
1983     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1984        sys/time.h, which is documented as needed for this function. It doesn't
1985        seem to matter on Linux, but it showed up on some releases of OS X.
1986    
1987     3. It seems that there are systems where bytes whose values are greater than
1988        127 match isprint() in the "C" locale. The "C" locale should be the
1989        default when a C program starts up. In most systems, only ASCII printing
1990        characters match isprint(). This difference caused the output from pcretest
1991        to vary, making some of the tests fail. I have changed pcretest so that:
1992    
1993        (a) When it is outputting text in the compiled version of a pattern, bytes
1994            other than 32-126 are always shown as hex escapes.
1995    
1996        (b) When it is outputting text that is a matched part of a subject string,
1997            it does the same, unless a different locale has been set for the match
1998            (using the /L modifier). In this case, it uses isprint() to decide.
1999    
2000     4. Fixed a major bug that caused incorrect computation of the amount of memory
2001        required for a compiled pattern when options that changed within the
2002        pattern affected the logic of the preliminary scan that determines the
2003        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2004        that the computed length was too small. The symptoms of this bug were
2005        either the PCRE error "internal error: code overflow" from pcre_compile(),
2006        or a glibc crash with a message such as "pcretest: free(): invalid next
2007        size (fast)". Examples of patterns that provoked this bug (shown in
2008        pcretest format) are:
2009    
2010          /(?-x: )/x
2011          /(?x)(?-x: \s*#\s*)/
2012          /((?i)[\x{c0}])/8
2013          /(?i:[\x{c0}])/8
2014    
2015        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2016        is now done differently.
2017    
2018     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2019        wrapper classes; (b) implement a new function in the C++ scanner that is
2020        more efficient than the old way of doing things because it avoids levels of
2021        recursion in the regex matching; (c) add a paragraph to the documentation
2022        for the FullMatch() function.
2023    
2024     6. The escape sequence \n was being treated as whatever was defined as
2025        "newline". Not only was this contrary to the documentation, which states
2026        that \n is character 10 (hex 0A), but it also went horribly wrong when
2027        "newline" was defined as CRLF. This has been fixed.
2028    
2029     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2030        was being set to -1 for the "end of line" case (supposedly a value that no
2031        character can have). Though this value is never used (the check for end of
2032        line is "zero bytes in current character"), it caused compiler complaints.
2033        I've changed it to 0xffffffff.
2034    
2035     8. In pcre_version.c, the version string was being built by a sequence of
2036        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2037        string (as it is for production releases) called a macro with an empty
2038        argument. The C standard says the result of this is undefined. The gcc
2039        compiler treats it as an empty string (which was what was wanted) but it is
2040        reported that Visual C gives an error. The source has been hacked around to
2041        avoid this problem.
2042    
2043     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2044        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2045        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2046        of them did).
2047    
2048    10. Originally, pcretest opened its input and output without "b"; then I was
2049        told that "b" was needed in some environments, so it was added for release
2050        5.0 to both the input and output. (It makes no difference on Unix-like
2051        systems.) Later I was told that it is wrong for the input on Windows. I've
2052        now abstracted the modes into two macros, to make it easier to fiddle with
2053        them, and removed "b" from the input mode under Windows.
2054    
2055    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2056    
2057    12. Added -help and --help to pcretest as an official way of being reminded
2058        of the options.
2059    
2060    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2061        and pcrecpp.cc because they annoy compilers at high warning levels.
2062    
2063    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2064    
2065    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2066        scripts are not C programs :-) and which was not noticed because it works
2067        on Linux.
2068    
2069    16. pcretest is supposed to handle any length of pattern and data line (as one
2070        line or as a continued sequence of lines) by extending its input buffer if
2071        necessary. This feature was broken for very long pattern lines, leading to
2072        a string of junk being passed to pcre_compile() if the pattern was longer
2073        than about 50K.
2074    
2075    17. I have done a major re-factoring of the way pcre_compile() computes the
2076        amount of memory needed for a compiled pattern. Previously, there was code
2077        that made a preliminary scan of the pattern in order to do this. That was
2078        OK when PCRE was new, but as the facilities have expanded, it has become
2079        harder and harder to keep it in step with the real compile phase, and there
2080        have been a number of bugs (see for example, 4 above). I have now found a
2081        cunning way of running the real compile function in a "fake" mode that
2082        enables it to compute how much memory it would need, while actually only
2083        ever using a few hundred bytes of working memory and without too many
2084        tests of the mode. This should make future maintenance and development
2085        easier. A side effect of this work is that the limit of 200 on the nesting
2086        depth of parentheses has been removed (though this was never a serious
2087        limitation, I suspect). However, there is a downside: pcre_compile() now
2088        runs more slowly than before (30% or more, depending on the pattern). I
2089        hope this isn't a big issue. There is no effect on runtime performance.
2090    
2091    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2092        newline (only possible for the last line of a file) and it was a
2093        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2094    
2095    19. Added additional timing features to pcretest. (1) The -tm option now times
2096        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2097        separate command line item, by a number that specifies the number of
2098        repeats to use when timing. The default is 50000; this gives better
2099        precision, but takes uncomfortably long for very large patterns.
2100    
2101    20. Extended pcre_study() to be more clever in cases where a branch of a
2102        subpattern has no definite first character. For example, (a*|b*)[cd] would
2103        previously give no result from pcre_study(). Now it recognizes that the
2104        first character must be a, b, c, or d.
2105    
2106    21. There was an incorrect error "recursive call could loop indefinitely" if
2107        a subpattern (or the entire pattern) that was being tested for matching an
2108        empty string contained only one non-empty item after a nested subpattern.
2109        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2110        incorrectly, because the \d was being skipped in the check.
2111    
2112    22. The pcretest program now has a new pattern option /B and a command line
2113        option -b, which is equivalent to adding /B to every pattern. This causes
2114        it to show the compiled bytecode, without the additional information that
2115        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2116        is the same as /B/I).
2117    
2118    23. A new optimization is now able automatically to treat some sequences such
2119        as a*b as a*+b. More specifically, if something simple (such as a character
2120        or a simple class like \d) has an unlimited quantifier, and is followed by
2121        something that cannot possibly match the quantified thing, the quantifier
2122        is automatically "possessified".
2123    
2124    24. A recursive reference to a subpattern whose number was greater than 39
2125        went wrong under certain circumstances in UTF-8 mode. This bug could also
2126        have affected the operation of pcre_study().
2127    
2128    25. Realized that a little bit of performance could be had by replacing
2129        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2130    
2131    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2132    
2133    27. Possessive quantifiers such as a++ were previously implemented by turning
2134        them into atomic groups such as ($>a+). Now they have their own opcodes,
2135        which improves performance. This includes the automatically created ones
2136        from 23 above.
2137    
2138    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2139        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2140        the first matched character to be a colon. This applied both to named and
2141        numbered groups.
2142    
2143    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2144    
2145    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2146        building PCRE on Windows easier, so I have included it in the distribution.
2147    
2148    31. There is now a check in pcretest against a ridiculously large number being
2149        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2150        loop, the loop is abandoned.
2151    
2152    32. Forward references to subpatterns in conditions such as (?(2)...) where
2153        subpattern 2 is defined later cause pcre_compile() to search forwards in
2154        the pattern for the relevant set of parentheses. This search went wrong
2155        when there were unescaped parentheses in a character class, parentheses
2156        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2157    
2158    33. "Subroutine" calls and backreferences were previously restricted to
2159        referencing subpatterns earlier in the regex. This restriction has now
2160        been removed.
2161    
2162    34. Added a number of extra features that are going to be in Perl 5.10. On the
2163        whole, these are just syntactic alternatives for features that PCRE had
2164        previously implemented using the Python syntax or my own invention. The
2165        other formats are all retained for compatibility.
2166    
2167        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2168            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2169            also .NET compatible.
2170    
2171        (b) A recursion or subroutine call to a named group can now be defined as
2172            (?&name) as well as (?P>name).
2173    
2174        (c) A backreference to a named group can now be defined as \k<name> or
2175            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2176            5.10, are also .NET compatible.
2177    
2178        (d) A conditional reference to a named group can now use the syntax
2179            (?(<name>) or (?('name') as well as (?(name).
2180    
2181        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2182            groups (named and numbered) that are never evaluated inline, but can be
2183            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2184            is always false. There may be only one alternative in such a group.
2185    
2186        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2187            as the simple (?(R). The condition is true only if the most recent
2188            recursion is that of the given number or name. It does not search out
2189            through the entire recursion stack.
2190    
2191        (g) The escape \gN or \g{N} has been added, where N is a positive or
2192            negative number, specifying an absolute or relative reference.
2193    
2194    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2195        some "unreachable code" warnings.
2196    
2197    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2198        things, this adds five new scripts.
2199    
2200    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2201        There were also incompatibilities regarding the handling of \Q..\E inside
2202        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2203        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2204    
2205    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2206        matches an empty string, and forcibly breaks the loop. There were bugs in
2207        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2208        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2209        separate and independent bugs (that affected different cases) have been
2210        fixed.
2211    
2212    39. Refactored the code to abolish the use of different opcodes for small
2213        capturing bracket numbers. This is a tidy that I avoided doing when I
2214        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2215        The new approach is not only tidier, it makes it possible to reduce the
2216        memory needed to fix the previous bug (38).
2217    
2218    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2219        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2220        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2221        mode.
2222    
2223    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2224        report.
2225    
2226    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2227        copy construction and assignment in the C++ wrapper.
2228    
2229    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2230        couple of bugs that could have given wrong results in the "--newline=crlf"
2231        case.
2232    
2233    44. Added a number of casts and did some reorganization of signed/unsigned int
2234        variables following suggestions from Dair Grant. Also renamed the variable
2235        "this" as "item" because it is a C++ keyword.
2236    
2237    45. Arranged for dftables to add
2238    
2239          #include "pcre_internal.h"
2240    
2241        to pcre_chartables.c because without it, gcc 4.x may remove the array
2242        definition from the final binary if PCRE is built into a static library and
2243        dead code stripping is activated.
2244    
2245    46. For an unanchored pattern, if a match attempt fails at the start of a
2246        newline sequence, and the newline setting is CRLF or ANY, and the next two
2247        characters are CRLF, advance by two characters instead of one.
2248    
2249    
2250    Version 6.7 04-Jul-06
2251    ---------------------
2252    
2253     1. In order to handle tests when input lines are enormously long, pcretest has
2254        been re-factored so that it automatically extends its buffers when
2255        necessary. The code is crude, but this _is_ just a test program. The
2256        default size has been increased from 32K to 50K.
2257    
2258     2. The code in pcre_study() was using the value of the re argument before
2259        testing it for NULL. (Of course, in any sensible call of the function, it
2260        won't be NULL.)
2261    
2262     3. The memmove() emulation function in pcre_internal.h, which is used on
2263        systems that lack both memmove() and bcopy() - that is, hardly ever -
2264        was missing a "static" storage class specifier.
2265    
2266     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2267        containing an extended class (one that cannot be represented by a bitmap
2268        because it contains high-valued characters or Unicode property items, e.g.
2269        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2270        pattern, but PCRE should not loop if you do not (it no longer does).
2271        [Detail: two cases were found: (a) a repeated subpattern containing an
2272        extended class; (b) a recursive reference to a subpattern that followed a
2273        previous extended class. It wasn't skipping over the extended class
2274        correctly when UTF-8 mode was not set.]
2275    
2276     5. A negated single-character class was not being recognized as fixed-length
2277        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2278        compile error "lookbehind assertion is not fixed length".
2279    
2280     6. The RunPerlTest auxiliary script was showing an unexpected difference
2281        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2282        write a Perl script that can interpret lines of an input file either as
2283        byte characters or as UTF-8, which is what "perltest" was being required to
2284        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2285        can't do is switch easily at run time between having the "use utf8;" pragma
2286        or not. In the end, I fudged it by using the RunPerlTest script to insert
2287        "use utf8;" explicitly for the UTF-8 tests.
2288    
2289     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2290        the end of the subject string, contrary to the documentation and to what
2291        Perl does. This was true of both matching functions. Now it matches only at
2292        the start of the subject and immediately after *internal* newlines.
2293    
2294     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2295        a pointer to an int instead of a pointer to an unsigned long int. This
2296        caused problems on 64-bit systems.
2297    
2298     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2299        instance of the 'standard' template library not being so standard".
2300    
2301    10. There was no check on the number of named subpatterns nor the maximum
2302        length of a subpattern name. The product of these values is used to compute
2303        the size of the memory block for a compiled pattern. By supplying a very
2304        long subpattern name and a large number of named subpatterns, the size
2305        computation could be caused to overflow. This is now prevented by limiting
2306        the length of names to 32 characters, and the number of named subpatterns
2307        to 10,000.
2308    
2309    11. Subpatterns that are repeated with specific counts have to be replicated in
2310        the compiled pattern. The size of memory for this was computed from the
2311        length of the subpattern and the repeat count. The latter is limited to
2312        65535, but there was no limit on the former, meaning that integer overflow
2313        could in principle occur. The compiled length of a repeated subpattern is
2314        now limited to 30,000 bytes in order to prevent this.
2315    
2316    12. Added the optional facility to have named substrings with the same name.
2317    
2318    13. Added the ability to use a named substring as a condition, using the
2319        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2320        are numbers (not recommended). Forward references are permitted.
2321    
2322    14. Added forward references in named backreferences (if you see what I mean).
2323    
2324    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2325        pattern could run off the end of the subject. For example, the pattern
2326        "(?s)(.{1,5})"8 did this with the subject "ab".
2327    
2328    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2329        PCRE_CASELESS was set when matching characters that were quantified with ?
2330        or *.
2331    
2332    17. A character class other than a single negated character that had a minimum
2333        but no maximum quantifier - for example [ab]{6,} - was not handled
2334        correctly by pce_dfa_exec(). It would match only one character.
2335    
2336    18. A valid (though odd) pattern that looked like a POSIX character
2337        class but used an invalid character after [ (for example [[,abc,]]) caused
2338        pcre_compile() to give the error "Failed: internal error: code overflow" or
2339        in some cases to crash with a glibc free() error. This could even happen if
2340        the pattern terminated after [[ but there just happened to be a sequence of
2341        letters, a binary zero, and a closing ] in the memory that followed.
2342    
2343    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2344        over the years. Originally (before any Unicode support), just the bottom 8
2345        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2346        output from "man perlunicode" includes this:
2347    
2348          The regular expression compiler produces polymorphic opcodes.  That
2349          is, the pattern adapts to the data and automatically switches to
2350          the Unicode character scheme when presented with Unicode data--or
2351          instead uses a traditional byte scheme when presented with byte
2352          data.
2353    
2354        Sadly, a wide octal escape does not cause a switch, and in a string with
2355        no other multibyte characters, these octal escapes are treated as before.
2356        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2357        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2358        Unicode string.
2359    
2360        I have not perpetrated such confusion in PCRE. Up till now, it took just
2361        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2362        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2363        translate to the appropriate multibyte character.
2364    
2365    29. Applied some refactoring to reduce the number of warnings from Microsoft
2366        and Borland compilers. This has included removing the fudge introduced
2367        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2368        a warning about an unused variable.
2369    
2370    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2371        characters since release 4.0, because Perl (from release 5.004) does not.
2372        [Or at least, is documented not to: some releases seem to be in conflict
2373        with the documentation.] However, when a pattern was studied with
2374        pcre_study() and all its branches started with \s, PCRE still included VT
2375        as a possible starting character. Of course, this did no harm; it just
2376        caused an unnecessary match attempt.
2377    
2378    22. Removed a now-redundant internal flag bit that recorded the fact that case
2379        dependency changed within the pattern. This was once needed for "required
2380        byte" processing, but is no longer used. This recovers a now-scarce options
2381        bit. Also moved the least significant internal flag bit to the most-
2382        significant bit of the word, which was not previously used (hangover from
2383        the days when it was an int rather than a uint) to free up another bit for
2384        the future.
2385    
2386    23. Added support for CRLF line endings as well as CR and LF. As well as the
2387        default being selectable at build time, it can now be changed at runtime
2388        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2389        specify that it is scanning data with non-default line endings.
2390    
2391    24. Changed the definition of CXXLINK to make it agree with the definition of
2392        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2393    
2394    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2395        recursions. This makes a big different to stack usage for some patterns.
2396    
2397    26. If a subpattern containing a named recursion or subroutine reference such
2398        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2399        the space required for the compiled pattern went wrong and gave too small a
2400        value. Depending on the environment, this could lead to "Failed: internal
2401        error: code overflow at offset 49" or "glibc detected double free or
2402        corruption" errors.
2403    
2404    27. Applied patches from Google (a) to support the new newline modes and (b) to
2405        advance over multibyte UTF-8 characters in GlobalReplace.
2406    
2407    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2408        difference for some implementation of PCRE in some Windows version.
2409    
2410    29. Added some extra testing facilities to pcretest:
2411    
2412        \q<number>   in a data line sets the "match limit" value
2413        \Q<number>   in a data line sets the "match recursion limt" value
2414        -S <number>  sets the stack size, where <number> is in megabytes
2415    
2416        The -S option isn't available for Windows.
2417    
2418    
2419    Version 6.6 06-Feb-06
2420    ---------------------
2421    
2422     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2423        in pcreposix.h. I have copied the definition from pcre.h.
2424    
2425     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2426        because pcre.h is no longer a built file.
2427    
2428     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2429        not normally included in the compiled code.
2430    
2431    
2432    Version 6.5 01-Feb-06
2433    ---------------------
2434    
2435     1. When using the partial match feature with pcre_dfa_exec(), it was not
2436        anchoring the second and subsequent partial matches at the new starting
2437        point. This could lead to incorrect results. For example, with the pattern
2438        /1234/, partially matching against "123" and then "a4" gave a match.
2439    
2440     2. Changes to pcregrep:
2441    
2442        (a) All non-match returns from pcre_exec() were being treated as failures
2443            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2444            error message is output. Some extra information is given for the
2445            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2446            probably the only errors that are likely to be caused by users (by
2447            specifying a regex that has nested indefinite repeats, for instance).
2448            If there are more than 20 of these errors, pcregrep is abandoned.
2449    
2450        (b) A binary zero was treated as data while matching, but terminated the
2451            output line if it was written out. This has been fixed: binary zeroes
2452            are now no different to any other data bytes.
2453    
2454        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2455            used to set a locale for matching. The --locale=xxxx long option has
2456            been added (no short equivalent) to specify a locale explicitly on the
2457            pcregrep command, overriding the environment variables.
2458    
2459        (d) When -B was used with -n, some line numbers in the output were one less
2460            than they should have been.
2461    
2462        (e) Added the -o (--only-matching) option.
2463    
2464        (f) If -A or -C was used with -c (count only), some lines of context were
2465            accidentally printed for the final match.
2466    
2467        (g) Added the -H (--with-filename) option.
2468    
2469        (h) The combination of options -rh failed to suppress file names for files
2470            that were found from directory arguments.
2471    
2472        (i) Added the -D (--devices) and -d (--directories) options.
2473    
2474        (j) Added the -F (--fixed-strings) option.
2475    
2476        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2477    
2478        (l) Added the --colo(u)r option.
2479    
2480        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2481            is not present by default.
2482    
2483     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2484        items such as (?R) or (?1), when the recursion could match a number of
2485        alternatives. If it matched one of the alternatives, but subsequently,
2486        outside the recursion, there was a failure, the code tried to back up into
2487        the recursion. However, because of the way PCRE is implemented, this is not
2488        possible, and the result was an incorrect result from the match.
2489    
2490        In order to prevent this happening, the specification of recursion has
2491        been changed so that all such subpatterns are automatically treated as
2492        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2493    
2494     4. I had overlooked the fact that, in some locales, there are characters for
2495        which isalpha() is true but neither isupper() nor islower() are true. In
2496        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2497        and ordfeminine) are like this. This affected the treatment of \w and \W
2498        when they appeared in character classes, but not when they appeared outside
2499        a character class. The bit map for "word" characters is now created
2500        separately from the results of isalnum() instead of just taking it from the
2501        upper, lower, and digit maps. (Plus the underscore character, of course.)
2502    
2503     5. The above bug also affected the handling of POSIX character classes such as
2504        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2505        permanent tables. Instead, the bit maps for such a class were previously
2506        created as the appropriate unions of the upper, lower, and digit bitmaps.
2507        Now they are created by subtraction from the [[:word:]] class, which has
2508        its own bitmap.
2509    
2510     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2511        It is created by subtracting the vertical space characters (\x09, \x0a,
2512        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2513        subtraction was done in the overall bitmap for a character class, meaning
2514        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2515        be recognized. This bug has been fixed.
2516    
2517     7. Patches from the folks at Google:
2518    
2519          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2520          real life, but is still worth protecting against".
2521    
2522          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2523          regular expressions".
2524    
2525          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2526          have it.
2527    
2528          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2529          "configure" and the latter not, in order to fix a problem somebody had
2530          with compiling the Arg class on HP-UX.
2531    
2532          (e) Improve the error-handling of the C++ wrapper a little bit.
2533    
2534          (f) New tests for checking recursion limiting.
2535    
2536     8. The pcre_memmove() function, which is used only if the environment does not
2537        have a standard memmove() function (and is therefore rarely compiled),
2538        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2539        returning a result (though PCRE never actually uses the result).
2540    
2541     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2542        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2543        returned instead of calling malloc() with an overflowing number that would
2544        most likely cause subsequent chaos.
2545    
2546    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2547    
2548    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2549        with this option is matched, the nmatch and pmatch options of regexec() are
2550        ignored.
2551    
2552    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2553        provided in case anyone wants to the the POSIX interface with UTF-8
2554        strings.
2555    
2556    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2557        C++ linking (needed for some HP-UX environments).
2558    
2559    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2560        (unused parameter) and in the pcre_printint() function (omitted "default"
2561        switch label when the default is to do nothing).
2562    
2563    15. Added some code to make it possible, when PCRE is compiled as a C++
2564        library, to replace subject pointers for pcre_exec() with a smart pointer
2565        class, thus making it possible to process discontinuous strings.
2566    
2567    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2568        much the same function. They were added by different people who were trying
2569        to make PCRE easy to compile on non-Unix systems. It has been suggested
2570        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2571        for compiling on Windows systems. I have therefore replaced it with
2572        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2573        defaults to "extern" for C or "extern C" for C++, which works fine on
2574        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2575        SCOPE with something explicit in config.h. In addition:
2576    
2577        (a) pcreposix.h still had just "extern" instead of either of these macros;
2578            I have replaced it with PCRE_DATA_SCOPE.
2579    
2580        (b) Functions such as _pcre_xclass(), which are internal to the library,
2581            but external in the C sense, all had PCRE_EXPORT in their definitions.
2582            This is apparently wrong for the Windows case, so I have removed it.
2583            (It makes no difference on Unix-like systems.)
2584    
2585    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2586        of recursive calls to match(). This is different to MATCH_LIMIT because
2587        that limits the total number of calls to match(), not all of which increase
2588        the depth of recursion. Limiting the recursion depth limits the amount of
2589        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2590        when PCRE is compiled, and changed at run time. A patch from Google adds
2591        this functionality to the C++ interface.
2592    
2593    18. Changes to the handling of Unicode character properties:
2594    
2595        (a) Updated the table to Unicode 4.1.0.
2596    
2597        (b) Recognize characters that are not in the table as "Cn" (undefined).
2598    
2599        (c) I revised the way the table is implemented to a much improved format
2600            which includes recognition of ranges. It now supports the ranges that
2601            are defined in UnicodeData.txt, and it also amalgamates other
2602            characters into ranges. This has reduced the number of entries in the
2603            table from around 16,000 to around 3,000, thus reducing its size
2604            considerably. I realized I did not need to use a tree structure after
2605            all - a binary chop search is just as efficient. Having reduced the
2606            number of entries, I extended their size from 6 bytes to 8 bytes to
2607            allow for more data.
2608    
2609        (d) Added support for Unicode script names via properties such as \p{Han}.
2610    
2611    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2612        matching that character.
2613    
2614    20. When matching a repeated Unicode property with a minimum greater than zero,
2615        (for example \pL{2,}), PCRE could look past the end of the subject if it
2616        reached it while seeking the minimum number of characters. This could
2617        happen only if some of the characters were more than one byte long, because
2618        there is a check for at least the minimum number of bytes.
2619    
2620    21. Refactored the implementation of \p and \P so as to be more general, to
2621        allow for more different types of property in future. This has changed the
2622        compiled form incompatibly. Anybody with saved compiled patterns that use
2623        \p or \P will have to recompile them.
2624    
2625    22. Added "Any" and "L&" to the supported property types.
2626    
2627    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2628        but give a compile time error if the value is greater than 0xff.
2629    
2630    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2631        accidentally not being installed or uninstalled.
2632    
2633    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2634        made were to insert the current release number. This seemed silly, because
2635        it made things harder for people building PCRE on systems that don't run
2636        "configure". I have turned pcre.h into a distributed file, no longer built
2637        by "configure", with the version identification directly included. There is
2638        no longer a pcre.h.in file.
2639    
2640        However, this change necessitated a change to the pcre-config script as
2641        well. It is built from pcre-config.in, and one of the substitutions was the
2642        release number. I have updated configure.ac so that ./configure now finds
2643        the release number by grepping pcre.h.
2644    
2645    26. Added the ability to run the tests under valgrind.
2646    
2647    
2648    Version 6.4 05-Sep-05
2649    ---------------------
2650    
2651     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2652        "--" to be printed when multiple files were scanned, even when none of the
2653        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2654        consider it to be a bug, and have restored the previous behaviour.
2655    
2656     2. A couple of code tidies to get rid of compiler warnings.
2657    
2658     3. The pcretest program used to cheat by referring to symbols in the library
2659        whose names begin with _pcre_. These are internal symbols that are not
2660        really supposed to be visible externally, and in some environments it is
2661        possible to suppress them. The cheating is now confined to including
2662        certain files from the library's source, which is a bit cleaner.
2663    
2664     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2665        file's purpose clearer.
2666    
2667     5. Reorganized pcre_ucp_findchar().
2668    
2669    
2670    Version 6.3 15-Aug-05
2671    ---------------------
2672    
2673     1. The file libpcre.pc.in did not have general read permission in the tarball.
2674    
2675     2. There were some problems when building without C++ support:
2676    
2677        (a) If C++ support was not built, "make install" and "make test" still
2678            tried to test it.
2679    
2680        (b) There were problems when the value of CXX was explicitly set. Some
2681            changes have been made to try to fix these, and ...
2682    
2683        (c) --disable-cpp can now be used to explicitly disable C++ support.
2684    
2685        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2686            backslash in a target when C++ was disabled. This confuses some
2687            versions of "make", apparently. Using an intermediate variable solves
2688            this. (Same for CPP_LOBJ.)
2689    
2690     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2691        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2692        necessary on certain architectures.
2693    
2694     4. Added a setting of -export-symbols-regex to the link command to remove
2695        those symbols that are exported in the C sense, but actually are local
2696        within the library, and not documented. Their names all begin with
2697        "_pcre_". This is not a perfect job, because (a) we have to except some
2698        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2699        available (and never for static libraries). I have made a note to try to
2700        find a way round (a) in the future.
2701    
2702    
2703    Version 6.2 01-Aug-05
2704    ---------------------
2705    
2706     1. There was no test for integer overflow of quantifier values. A construction
2707        such as {1111111111111111} would give undefined results. What is worse, if
2708        a minimum quantifier for a parenthesized subpattern overflowed and became
2709        negative, the calculation of the memory size went wrong. This could have
2710        led to memory overwriting.
2711    
2712     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2713    
2714     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2715        operating environments where this matters.
2716    
2717     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2718        PCRE options from within the C++ wrapper.
2719    
2720     5. Named capturing subpatterns were not being correctly counted when a pattern
2721        was compiled. This caused two problems: (a) If there were more than 100
2722        such subpatterns, the calculation of the memory needed for the whole
2723        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2724        back references of the form \12, where the number was greater than 9, were
2725        not recognized as back references, even though there were sufficient
2726        previous subpatterns.
2727    
2728     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2729        versions of gcc, e.g. 2.95.4.
2730    
2731    
2732    Version 6.1 21-Jun-05
2733    ---------------------
2734    
2735     1. There was one reference to the variable "posix" in pcretest.c that was not
2736        surrounded by "#if !defined NOPOSIX".
2737    
2738     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2739        the cross-check on the old pcre_info() function, for the benefit of the
2740        cut-down version of PCRE that is currently imported into Exim.
2741    
2742     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2743        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2744        patterns that start (?i) but I don't think that matters. The use of (?i) is
2745        just an example; this all applies to the other options as well.
2746    
2747     4. Since libtool seems to echo the compile commands it is issuing, the output
2748        from "make" can be reduced a bit by putting "@" in front of each libtool
2749        compile command.
2750    
2751     5. Patch from the folks at Google for configure.in to be a bit more thorough
2752        in checking for a suitable C++ installation before trying to compile the
2753        C++ stuff. This should fix a reported problem when a compiler was present,
2754        but no suitable headers.
2755    
2756     6. The man pages all had just "PCRE" as their title. I have changed them to
2757        be the relevant file name. I have also arranged that these names are
2758        retained in the file doc/pcre.txt, which is a concatenation in text format
2759        of all the man pages except the little individual ones for each function.
2760    
2761     7. The NON-UNIX-USE file had not been updated for the different set of source
2762        files that come with release 6. I also added a few comments about the C++
2763        wrapper.
2764    
2765    
2766    Version 6.0 07-Jun-05
2767    ---------------------
2768    
2769     1. Some minor internal re-organization to help with my DFA experiments.
2770    
2771     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2772        didn't matter for the library itself when fully configured, but did matter
2773        when compiling without UCP support, or within Exim, where the ucp files are
2774        not imported.
2775    
2776     3. Refactoring of the library code to split up the various functions into
2777        different source modules. The addition of the new DFA matching code (see
2778        below) to a single monolithic source would have made it really too
2779        unwieldy, quite apart from causing all the code to be include in a
2780        statically linked application, when only some functions are used. This is
2781        relevant even without the DFA addition now that patterns can be compiled in
2782        one application and matched in another.
2783    
2784        The downside of splitting up is that there have to be some external
2785        functions and data tables that are used internally in different modules of
2786        the library but which are not part of the API. These have all had their
2787        names changed to start with "_pcre_" so that they are unlikely to clash
2788        with other external names.
2789    
2790     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2791        a different (DFA) algorithm. Although it is slower than the original
2792        function, it does have some advantages for certain types of matching
2793        problem.
2794    
2795     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2796        including restarting after a partial match.
2797    
2798     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2799        defined when compiling for Windows was sent to me. I have put it into the
2800        code, though I have no means of testing or verifying it.
2801    
2802     7. Added the pcre_refcount() auxiliary function.
2803    
2804     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2805        match before or at the first newline in the subject string. In pcretest,
2806        the /f option on a pattern can be used to set this.
2807    
2808     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2809        would behave wrongly. This has been present in PCRE since release 4.0.
2810    
2811    10. A number of changes to the pcregrep command:
2812    
2813        (a) Refactored how -x works; insert ^(...)$ instead of setting
2814            PCRE_ANCHORED and checking the length, in preparation for adding
2815            something similar for -w.
2816    
2817        (b) Added the -w (match as a word) option.
2818    
2819        (c) Refactored the way lines are read and buffered so as to have more
2820            than one at a time available.
2821    
2822        (d) Implemented a pcregrep test script.
2823    
2824        (e) Added the -M (multiline match) option. This allows patterns to match
2825            over several lines of the subject. The buffering ensures that at least
2826            8K, or the rest of the document (whichever is the shorter) is available
2827            for matching (and similarly the previous 8K for lookbehind assertions).
2828    
2829        (f) Changed the --help output so that it now says
2830    
2831              -w, --word-regex(p)
2832    
2833            instead of two lines, one with "regex" and the other with "regexp"
2834            because that confused at least one person since the short forms are the
2835            same. (This required a bit of code, as the output is generated
2836            automatically from a table. It wasn't just a text change.)
2837    
2838        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2839            option but starts with a hyphen. Could be a pattern or a path name
2840            starting with a hyphen, for instance.
2841    
2842        (h) "-" can be given as a file name to represent stdin.
2843    
2844        (i) When file names are being printed, "(standard input)" is used for
2845            the standard input, for compatibility with GNU grep. Previously
2846            "<stdin>" was used.
2847    
2848        (j) The option --label=xxx can be used to supply a name to be used for
2849            stdin when file names are being printed. There is no short form.
2850    
2851        (k) Re-factored the options decoding logic because we are going to add
2852            two more options that take data. Such options can now be given in four
2853            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2854    
2855        (l) Added the -A, -B, and -C options for requesting that lines of context
2856            around matches be printed.
2857    
2858        (m) Added the -L option to print the names of files that do not contain
2859            any matching lines, that is, the complement of -l.
2860    
2861        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2862            continue to scan other files.
2863    
2864        (o) The -s option was incorrectly implemented. For compatibility with other
2865            greps, it now suppresses the error message for a non-existent or non-
2866            accessible file (but not the return code). There is a new option called
2867            -q that suppresses the output of matching lines, which was what -s was
2868            previously doing.
2869    
2870        (p) Added --include and --exclude options to specify files for inclusion
2871            and exclusion when recursing.
2872    
2873    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2874        Hopefully, it now does.
2875    
2876    12. Missing cast in pcre_study().
2877    
2878    13. Added an "uninstall" target to the makefile.
2879    
2880    14. Replaced "extern" in the function prototypes in Makefile.in with
2881        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2882        world, but is set differently for Windows.
2883    
2884    15. Added a second compiling function called pcre_compile2(). The only
2885        difference is that it has an extra argument, which is a pointer to an
2886        integer error code. When there is a compile-time failure, this is set
2887        non-zero, in addition to the error test pointer being set to point to an
2888        error message. The new argument may be NULL if no error number is required
2889        (but then you may as well call pcre_compile(), which is now just a
2890        wrapper). This facility is provided because some applications need a
2891        numeric error indication, but it has also enabled me to tidy up the way
2892        compile-time errors are handled in the POSIX wrapper.
2893    
2894    16. Added VPATH=.libs to the makefile; this should help when building with one
2895        prefix path and installing with another. (Or so I'm told by someone who
2896        knows more about this stuff than I do.)
2897    
2898    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2899        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2900        match everything, including newlines. This is not POSIX-compatible, but
2901        somebody wanted the feature. From pcretest it can be activated by using
2902        both the P and the s flags.
2903    
2904    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2905    
2906    19. libpcre.pc was being incorrectly installed as executable.
2907    
2908    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2909        it now also looks for '\r' so that it will work unmodified on Windows.
2910    
2911    21. Added Google's contributed C++ wrapper to the distribution.
2912    
2913    22. Added some untidy missing memory free() calls in pcretest, to keep
2914        Electric Fence happy when testing.
2915    
2916    
2917    
2918    Version 5.0 13-Sep-04
2919    ---------------------
2920    
2921     1. Internal change: literal characters are no longer packed up into items
2922        containing multiple characters in a single byte-string. Each character
2923        is now matched using a separate opcode. However, there may be more than one
2924        byte in the character in UTF-8 mode.
2925    
2926     2. The pcre_callout_block structure has two new fields: pattern_position and
2927        next_item_length. These contain the offset in the pattern to the next match
2928        item, and its length, respectively.
2929    
2930     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2931        insertion of callouts before each pattern item. Added the /C option to
2932        pcretest to make use of this.
2933    
2934     4. On the advice of a Windows user, the lines
2935    
2936          #if defined(_WIN32) || defined(WIN32)
2937          _setmode( _fileno( stdout ), 0x8000 );
2938          #endif  /* defined(_WIN32) || defined(WIN32) */
2939    
2940        have been added to the source of pcretest. This apparently does useful
2941        magic in relation to line terminators.
2942    
2943     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2944        for the benefit of those environments where the "b" makes a difference.
2945    
2946     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2947        to know about it. I have put a hack into configure.in that adds in code
2948        to set GCC=yes if CC=icc. This seems to end up at a point in the
2949        generated configure script that is early enough to affect the setting of
2950        compiler options, which is what is needed, but I have no means of testing
2951        whether it really works. (The user who reported this had patched the
2952        generated configure script, which of course I cannot do.)
2953    
2954        LATER: After change 22 below (new libtool files), the configure script
2955        seems to know about icc (and also ecc). Therefore, I have commented out
2956        this hack in configure.in.
2957    
2958     7. Added support for pkg-config (2 patches were sent in).
2959    
2960     8. Negated POSIX character classes that used a combination of internal tables
2961        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2962        [[:^ascii]]. Typically, they would match almost any characters. The other
2963        POSIX classes were not broken in this way.
2964    
2965     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2966        to find the match, as PCRE was deluded into thinking that the match had to
2967        start at the start point or following a newline. The same bug applied to
2968        patterns with negative forward assertions or any backward assertions
2969        preceding ".*" at the start, unless the pattern required a fixed first
2970        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2971    
2972    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2973        starting at the last subject character, bytes beyond the end of the subject
2974        string were read.
2975    
2976    11. Renamed the variable "class" as "classbits" to make life easier for C++
2977        users. (Previously there was a macro definition, but it apparently wasn't
2978        enough.)
2979    
2980    12. Added the new field "tables" to the extra data so that tables can be passed
2981        in at exec time, or the internal tables can be re-selected. This allows
2982        a compiled regex to be saved and re-used at a later time by a different
2983        program that might have everything at different addresses.
2984    
2985    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2986        -R library as well as a -L library.
2987    
2988    14. The debugging options of pcretest (-d on the command line or D on a
2989        pattern) showed incorrect output for anything following an extended class
2990        that contained multibyte characters and which was followed by a quantifier.
2991    
2992    15. Added optional support for general category Unicode character properties
2993        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2994        support. It adds about 90K to the size of the library. The meanings of the
2995        inbuilt class escapes such as \d and \s have NOT been changed.
2996    
2997    16. Updated pcredemo.c to include calls to free() to release the memory for the
2998        compiled pattern.
2999    
3000    17. The generated file chartables.c was being created in the source directory
3001        instead of in the building directory. This caused the build to fail if the
3002        source directory was different from the building directory, and was
3003        read-only.
3004    
3005    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3006        file. No doubt somebody will tell me if they don't make sense... Also added
3007        Dan Mooney's comments about building on OpenVMS.
3008    
3009    19. Added support for partial matching via the PCRE_PARTIAL option for
3010        pcre_exec() and the \P data escape in pcretest.
3011    
3012    20. Extended pcretest with 3 new pattern features:
3013    
3014        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3015              write the compiled pattern to the file whose name is "rest-of-line".
3016              This is a straight binary dump of the data, with the saved pointer to
3017              the character tables forced to be NULL. The study data, if any, is
3018              written too. After writing, pcretest reads a new pattern.
3019    
3020        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3021              compiled pattern from the given file. There must not be any
3022              occurrences of "<" in the file name (pretty unlikely); if there are,
3023              pcretest will instead treat the initial "<" as a pattern delimiter.
3024              After reading in the pattern, pcretest goes on to read data lines as
3025              usual.
3026    
3027        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3028              and 16-bit fields in a compiled pattern, to simulate a pattern that
3029              was compiled on a host of opposite endianness.
3030    
3031    21. The pcre-exec() function can now cope with patterns that were compiled on
3032        hosts of opposite endianness, with this restriction:
3033    
3034          As for any compiled expression that is saved and used later, the tables
3035          pointer field cannot be preserved; the extra_data field in the arguments
3036          to pcre_exec() should be used to pass in a tables address if a value
3037          other than the default internal tables were used at compile time.
3038    
3039    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3040        now diagnosed as an error. Previously, most of the time, a negative number
3041        would have been treated as zero, but if in addition "ovector" was passed as
3042        NULL, a crash could occur.
3043    
3044    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3045        new versions from the libtool 1.5 distribution (the last one is a copy of
3046        a file called libtool.m4). This seems to have fixed the need to patch
3047        "configure" to support Darwin 1.3 (which I used to do). However, I still
3048        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3049        workstation).
3050    
3051    24. Changed the PCRE licence to be the more standard "BSD" licence.
3052    
3053    
3054    Version 4.5 01-Dec-03
3055    ---------------------
3056    
3057     1. There has been some re-arrangement of the code for the match() function so
3058        that it can be compiled in a version that does not call itself recursively.
3059        Instead, it keeps those local variables that need separate instances for
3060        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3061        needs to "recurse". Keeping track of where control must go is done by means
3062        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3063        hide most of the details from the main code, and operates only if
3064        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3065        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3066        operating.
3067    
3068        To make it easier for callers to provide specially tailored get/free
3069        functions for this usage, two new functions, pcre_stack_malloc, and
3070        pcre_stack_free, are used. They are always called in strict stacking order,
3071        and the size of block requested is always the same.
3072    
3073        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3074        PCRE has been compiled to use the stack or the heap for recursion. The
3075        -C option of pcretest uses this to show which version is compiled.
3076    
3077        A new data escape \S, is added to pcretest; it causes the amounts of store
3078        obtained and freed by both kinds of malloc/free at match time to be added
3079        to the output.
3080    
3081     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3082        what's available on my current Linux desktop machine.
3083    
3084     3. When matching a UTF-8 string, the test for a valid string at the start has
3085        been extended. If start_offset is not zero, PCRE now checks that it points
3086        to a byte that is the start of a UTF-8 character. If not, it returns
3087        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3088        this is necessary because there may be backward assertions in the pattern.
3089        When matching the same subject several times, it may save resources to use
3090        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3091    
3092     4. The code for checking the validity of UTF-8 strings has been tightened so
3093        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3094        containing "overlong sequences".
3095    
3096     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3097        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3098        should have been "&", but it just so happened that all the cases this let
3099        through by mistake were picked up later in the function.
3100    
3101     6. I had used a variable called "isblank" - this is a C99 function, causing
3102        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3103    
3104     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3105        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3106        so the version is shown; (c) stop "make test" echoing "./RunTest".
3107    
3108     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3109    
3110     9. The prototype for memmove() for systems that don't have it was using
3111        size_t, but the inclusion of the header that defines size_t was later. I've
3112        moved the #includes for the C headers earlier to avoid this.
3113    
3114    10. Added some adjustments to the code to make it easier to compiler on certain
3115        special systems:
3116    
3117          (a) Some "const" qualifiers were missing.
3118          (b) Added the macro EXPORT before all exported functions; by default this
3119              is defined to be empty.
3120          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3121              that it reads its output file name as an argument instead of writing
3122              to the standard output and assuming this can be redirected.
3123    
3124    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3125        class containing characters with values greater than 255, PCRE compilation
3126        went into a loop.
3127    
3128    12. A recursive reference to a subpattern that was within another subpattern
3129        that had a minimum quantifier of zero caused PCRE to crash. For example,
3130        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3131        recursion. If the recursively-called subpattern itself had a zero repeat,
3132        that was OK.
3133    
3134    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3135        buffer into which it was copied (for escape processing) was still set at
3136        1024, so long lines caused crashes.
3137    
3138    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3139        "internal error: code overflow...". This applied to any character class
3140        that was followed by a possessive quantifier.
3141    
3142    15. Modified the Makefile to add libpcre.la as a prerequisite for
3143        libpcreposix.la because I was told this is needed for a parallel build to
3144        work.
3145    
3146    16. If a pattern that contained .* following optional items at the start was
3147        studied, the wrong optimizing data was generated, leading to matching
3148        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3149        matching string must start with a or b or c. The correct conclusion for
3150        this pattern is that a match can start with any character.
3151    
3152    
3153    Version 4.4 13-Aug-03
3154    ---------------------
3155    
3156     1. In UTF-8 mode, a character class containing characters with values between
3157        127 and 255 was not handled correctly if the compiled pattern was studied.
3158        In fixing this, I have also improved the studying algorithm for such
3159        classes (slightly).
3160    
3161     2. Three internal functions had redundant arguments passed to them. Removal
3162        might give a very teeny performance improvement.
3163    
3164     3. Documentation bug: the value of the capture_top field in a callout is *one
3165        more than* the number of the hightest numbered captured substring.
3166    
3167     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3168        in incorrectly linking with a previously installed version. They now link
3169        explicitly with libpcre.la.
3170    
3171     5. configure.in no longer needs to recognize Cygwin specially.
3172    
3173     6. A problem in pcre.in for Windows platforms is fixed.
3174    
3175     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3176        pcretest, it used to include the size of the study block as part of its
3177        output. Unfortunately, the structure contains a field that has a different
3178        size on different hardware architectures. This meant that the tests that
3179        showed this size failed. As the block is currently always of a fixed size,
3180        this information isn't actually particularly useful in pcretest output, so
3181        I have just removed it.
3182    
3183     8. Three pre-processor statements accidentally did not start in column 1.
3184        Sadly, there are *still* compilers around that complain, even though
3185        standard C has not required this for well over a decade. Sigh.
3186    
3187     9. In pcretest, the code for checking callouts passed small integers in the
3188        callout_data field, which is a void * field. However, some picky compilers
3189        complained about the casts involved for this on 64-bit systems. Now
3190        pcretest passes the address of the small integer instead, which should get
3191        rid of the warnings.
3192    
3193    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3194        both compile and run time, and gives an error if an invalid UTF-8 sequence
3195        is found. There is a option for disabling this check in cases where the
3196        string is known to be correct and/or the maximum performance is wanted.
3197    
3198    11. In response to a bug report, I changed one line in Makefile.in from
3199    
3200            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3201        to
3202            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3203    
3204        to look similar to other lines, but I have no way of telling whether this
3205        is the right thing to do, as I do not use Windows. No doubt I'll get told
3206        if it's wrong...
3207    
3208    
3209    Version 4.3 21-May-03
3210    ---------------------
3211    
3212    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3213       Makefile.
3214    
3215    2. Some refactoring to improve the quality of the code:
3216    
3217       (i)   The utf8_table... variables are now declared "const".
3218    
3219       (ii)  The code for \cx, which used the "case flipping" table to upper case
3220             lower case letters, now just substracts 32. This is ASCII-specific,
3221             but the whole concept of \cx is ASCII-specific, so it seems
3222             reasonable.
3223    
3224       (iii) PCRE was using its character types table to recognize decimal and
3225             hexadecimal digits in the pattern. This is silly, because it handles
3226             only 0-9, a-f, and A-F, but the character types table is locale-
3227             specific, which means strange things might happen. A private
3228             table is now used for this - though it costs 256 bytes, a table is
3229             much faster than multiple explicit tests. Of course, the standard
3230             character types table is still used for matching digits in subject
3231             strings against \d.
3232    
3233       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3234             ending in _t are). So I've renamed it as ESC_tee.
3235    
3236    3. The first argument for regexec() in the POSIX wrapper should have been
3237       defined as "const".
3238    
3239    4. Changed pcretest to use malloc() for its buffers so that they can be
3240       Electric Fenced for debugging.
3241    
3242    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3243       to read one or more bytes before the start of the subject string. Often this
3244       had no effect on PCRE's behaviour, but in some circumstances it could
3245       provoke a segmentation fault.
3246    
3247    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3248       to try to read one or more bytes before the start of the subject string.
3249    
3250    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3251       UTF-8 support could misbehave in various ways if the subject string
3252       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3253       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3254       back over UTF-8 characters.)
3255    
3256    
3257    Version 4.2 14-Apr-03
3258    ---------------------
3259    
3260    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3261    
3262    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3263         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3264         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3265         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3266         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3267           and BUILD_EXEEXT
3268         Note: automatic setting of the BUILD variables is not yet working
3269         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3270           compile-time but not at link-time
3271         [LINK]: use for linking executables only
3272         make different versions for Windows and non-Windows
3273         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3274           libraries
3275         [LINK_FOR_BUILD]: new variable
3276         [OBJEXT]: use throughout
3277         [EXEEXT]: use throughout
3278         <winshared>: new target
3279         <wininstall>: new target
3280         <dftables.o>: use native compiler
3281         <dftables>: use native linker
3282         <install>: handle Windows platform correctly
3283         <clean>: ditto
3284         <check>: ditto
3285         copy DLL to top builddir before testing
3286    
3287       As part of these changes, -no-undefined was removed again. This was reported
3288       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3289       in any case.
3290    
3291    3. Some tidies to get rid of compiler warnings:
3292    
3293       . In the match_data structure, match_limit was an unsigned long int, whereas
3294         match_call_count was an int. I've made them both unsigned long ints.
3295    
3296       . In pcretest the fact that a const uschar * doesn't automatically cast to
3297         a void * provoked a warning.
3298    
3299       . Turning on some more compiler warnings threw up some "shadow" variables
3300         and a few more missing casts.
3301    
3302    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3303       option, a class that contained a single character with a value between 128
3304       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3305    
3306    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3307       option, a class that contained several characters, but with at least one
3308       whose value was between 128 and 255 caused PCRE to crash.
3309    
3310    
3311    Version 4.1 12-Mar-03
3312    ---------------------
3313    
3314    1. Compiling with gcc -pedantic found a couple of places where casts were
3315    needed, and a string in dftables.c that was longer than standard compilers are
3316    required to support.
3317    
3318    2. Compiling with Sun's compiler found a few more places where the code could
3319    be tidied up in order to avoid warnings.
3320    
3321    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3322    first of these names is deprecated in the latest Autoconf in favour of the name
3323    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3324    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3325    analogy I have changed it to CFLAGS_FOR_BUILD.
3326    
3327    4. Added -no-undefined to the linking command in the Makefile, because this is
3328    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3329    linking step for the pcreposix library.
3330    
3331    5. PCRE was failing to diagnose the case of two named groups with the same
3332    name.
3333    
3334    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3335    literal character that is needed in the subject for a match, and scans along to
3336    ensure that it is present before embarking on the full matching process. This
3337    saves time in cases of nested unlimited repeats that are never going to match.
3338    Problem: the scan can take a lot of time if the subject is very long (e.g.
3339    megabytes), thus penalizing straightforward matches. It is now done only if the
3340    amount of subject to be scanned is less than 1000 bytes.
3341    
3342    7. A lesser problem with the same optimization is that it was recording the
3343    first character of an anchored pattern as "needed", thus provoking a search
3344    right along the subject, even when the first match of the pattern was going to
3345    fail. The "needed" character is now not set for anchored patterns, unless it
3346    follows something in the pattern that is of non-fixed length. Thus, it still
3347    fulfils its original purpose of finding quick non-matches in cases of nested
3348    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3349    
3350    
3351    Version 4.0 17-Feb-03
3352    ---------------------
3353    
3354    1. If a comment in an extended regex that started immediately after a meta-item
3355    extended to the end of string, PCRE compiled incorrect data. This could lead to
3356    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3357    
3358    2. Moved to autoconf 2.53 and libtool 1.4.2.
3359    
3360    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3361    the special perltest8 script is no longer needed - all the tests can be run
3362    from a single perltest script.
3363    
3364    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3365    by \s. It has now been removed in PCRE. This means it isn't recognized as
3366    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3367    class [:space:] *does* include VT, thereby creating a mess.
3368    
3369    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3370    space and tab.
3371    
3372    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3373    its new features into the main test script, reducing the number of scripts.
3374    
3375    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3376    were backward compatible, and made the (?i) apply to the whole pattern, as if
3377    /i were given. Now it behaves more logically, and applies the option setting
3378    only to what follows. PCRE has been changed to follow suit. However, if it
3379    finds options settings right at the start of the pattern, it extracts them into
3380    the global options, as before. Thus, they show up in the info data.
3381    
3382    8. Added support for the \Q...\E escape sequence. Characters in between are
3383    treated as literals. This is slightly different from Perl in that $ and @ are
3384    also handled as literals inside the quotes. In Perl, they will cause variable
3385    interpolation. Note the following examples:
3386    
3387        Pattern            PCRE matches      Perl matches
3388    
3389        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3390        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3391        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3392    
3393    For compatibility with Perl, \Q...\E sequences are recognized inside character
3394    classes as well as outside them.
3395    
3396    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3397    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3398    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3399    signed/unsigned warnings.
3400    
3401    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3402    option for pcretest, so I've replaced it by a simple function that does just
3403    that job.
3404    
3405    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3406    "pcregrep -".
3407    
3408    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3409    Java package. This provides some syntactic sugar for simple cases of what my
3410    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3411    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3412    item, you can use this simplified notation. Note that only makes sense with
3413    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3414    greediness, whatever the setting of the PCRE_UNGREEDY option.
3415    
3416    13. A change of greediness default within a pattern was not taking effect at
3417    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3418    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3419    was abstracted outside.
3420    
3421    14. PCRE now supports the \G assertion. It is true when the current matching
3422    position is at the start point of the match. This differs from \A when the
3423    starting offset is non-zero. Used with the /g option of pcretest (or similar
3424    code), it works in the same way as it does for Perl's /g option. If all
3425    alternatives of a regex begin with \G, the expression is anchored to the start
3426    match position, and the "anchored" flag is set in the compiled expression.
3427    
3428    15. Some bugs concerning the handling of certain option changes within patterns
3429    have been fixed. These applied to options other than (?ims). For example,
3430    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3431    been the other way round. Some of this was related to change 7 above.
3432    
3433    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3434    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3435    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3436    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3437    
3438    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3439    mode. Unlike ".", it always matches newline, whatever the setting of
3440    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3441    assertions. Perl allows it, but it doesn't (in general) work because it can't
3442    calculate the length of the lookbehind. At least, that's the case for Perl
3443    5.8.0 - I've been told they are going to document that it doesn't work in
3444    future.
3445    
3446    18. Added an error diagnosis for escapes that PCRE does not support: these are
3447    \L, \l, \N, \P, \p, \U, \u, and \X.
3448    
3449    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3450    reading past the end of the pattern in cases such as /[abcd/.
3451    
3452    20. PCRE was getting more memory than necessary for patterns with classes that
3453    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3454    
3455    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3456    compiling PCRE for use with Virtual Pascal.
3457    
3458    22. Small fix to the Makefile to make it work properly if the build is done
3459    outside the source tree.
3460    
3461    23. Added a new extension: a condition to go with recursion. If a conditional
3462    subpattern starts with (?(R) the "true" branch is used if recursion has
3463    happened, whereas the "false" branch is used only at the top level.
3464    
3465    24. When there was a very long string of literal characters (over 255 bytes
3466    without UTF support, over 250 bytes with UTF support), the computation of how
3467    much memory was required could be incorrect, leading to segfaults or other
3468    strange effects.
3469    
3470    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3471    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3472    there was a subsequent back reference to those brackets. This meant that, for
3473    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3474    possible to check for precisely this case. All we can do is abandon the
3475    optimization if .* occurs inside capturing brackets when there are any back
3476    references whatsoever. (See below for a better fix that came later.)
3477    
3478    26. The handling of the optimization for finding the first character of a
3479    non-anchored pattern, and for finding a character that is required later in the
3480    match were failing in some cases. This didn't break the matching; it just
3481    failed to optimize when it could. The way this is done has been re-implemented.
3482    
3483    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3484    
3485    28. Added a new feature that provides some of the functionality that Perl
3486    provides with (?{...}). The facility is termed a "callout". The way it is done
3487    in PCRE is for the caller to provide an optional function, by setting
3488    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3489    global variable. By default it is unset, which disables all calling out. To get
3490    the function called, the regex must include (?C) at appropriate points. This
3491    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3492    This provides a means of identifying different callout points. When PCRE
3493    reaches such a point in the regex, if pcre_callout has been set, the external
3494    function is called. It is provided with data in a structure called
3495    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3496    matching continues; if it returns a non-zero value, the match at the current
3497    point fails. However, backtracking will occur if possible. [This was changed
3498    later and other features added - see item 49 below.]
3499    
3500    29. pcretest is upgraded to test the callout functionality. It provides a
3501    callout function that displays information. By default, it shows the start of
3502    the match and the current position in the text. There are some new data escapes
3503    to vary what happens:
3504    
3505        \C+         in addition, show current contents of captured substrings
3506        \C-         do not supply a callout function
3507        \C!n        return 1 when callout number n is reached
3508        \C!n!m      return 1 when callout number n is reached for the mth time
3509    
3510    30. If pcregrep was called with the -l option and just a single file name, it
3511    output "<stdin>" if a match was found, instead of the file name.
3512    
3513    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3514    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3515    pcre_exec(). This saves a malloc/free per call. The default value of
3516    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3517    when configuring.
3518    
3519    32. The default maximum size of a compiled pattern is 64K. There have been a
3520    few cases of people hitting this limit. The code now uses macros to handle the
3521    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3522    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3523    configuring. Tests 2 and 5 work only with 2-byte links because they output
3524    debugging information about compiled patterns.
3525    
3526    33. Internal code re-arrangements:
3527    
3528    (a) Moved the debugging function for printing out a compiled regex into
3529        its own source file (printint.c) and used #include to pull it into
3530        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3531        separate copies.
3532    
3533    (b) Defined the list of op-code names for debugging as a macro in
3534        internal.h so that it is next to the definition of the opcodes.
3535    
3536    (c) Defined a table of op-code lengths for simpler skipping along compiled
3537        code. This is again a macro in internal.h so that it is next to the
3538        definition of the opcodes.
3539    
3540    34. Added support for recursive calls to individual subpatterns, along the
3541    lines of Robin Houston's patch (but implemented somewhat differently).
3542    
3543    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3544    allow it to read and process whole directories in Win32. This code was
3545    contributed by Lionel Fourquaux; it has not been tested by me.
3546    
3547    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3548    used to name a group. Names consist of alphanumerics and underscores, and must
3549    be unique. Back references use the syntax (?P=name) and recursive calls use
3550    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3551    numbers. The function pcre_fullinfo() can be used after compilation to extract
3552    a name/number map. There are three relevant calls:
3553    
3554      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3555      PCRE_INFO_NAMECOUNT            yields the number of entries
3556      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3557    
3558    The map is a vector of fixed-size entries. The size of each entry depends on
3559    the length of the longest name used. The first two bytes of each entry are the
3560    group number, most significant byte first. There follows the corresponding
3561    name, zero terminated. The names are in alphabetical order.
3562    
3563    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3564    case instead of 255. Making it the same both with and without UTF-8 support
3565    means that the same test output works with both.
3566    
3567    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3568    calling malloc() with a zero argument.
3569    
3570    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3571    optimization. I've improved things by keeping a bitmap of backreferences with
3572    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3573    fact referenced, the optimization can be applied. It is unlikely that a
3574    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3575    the match to follow \n) will appear inside brackets with a number greater than
3576    31, but if it does, any back reference > 31 suppresses the optimization.
3577    
3578    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3579    of disabling numbered capturing parentheses. Any opening parenthesis that is
3580    not followed by ? behaves as if it were followed by ?: but named parentheses
3581    can still be used for capturing (and they will acquire numbers in the usual
3582    way).
3583    
3584    41. Redesigned the return codes from the match() function into yes/no/error so
3585    that errors can be passed back from deep inside the nested calls. A malloc
3586    failure while inside a recursive subpattern call now causes the
3587    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3588    
3589    42. It is now possible to set a limit on the number of times the match()
3590    function is called in a call to pcre_exec(). This facility makes it possible to
3591    limit the amount of recursion and backtracking, though not in a directly
3592    obvious way, because the match() function is used in a number of different
3593    circumstances. The count starts from zero for each position in the subject
3594    string (for non-anchored patterns). The default limit is, for compatibility, a
3595    large number, namely 10 000 000. You can change this in two ways:
3596    
3597    (a) When configuring PCRE before making, you can use --with-match-limit=n
3598        to set a default value for the compiled library.
3599    
3600    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3601        a different value is set. See 45 below.
3602    
3603    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3604    
3605    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3606    of things that can be changed at compile time. The first argument specifies
3607    what is wanted and the second points to where the information is to be placed.
3608    The current list of available information is:
3609    
3610      PCRE_CONFIG_UTF8
3611    
3612    The output is an integer that is set to one if UTF-8 support is available;
3613    otherwise it is set to zero.
3614    
3615      PCRE_CONFIG_NEWLINE
3616    
3617    The output is an integer that it set to the value of the code that is used for
3618    newline. It is either LF (10) or CR (13).
3619    
3620      PCRE_CONFIG_LINK_SIZE
3621    
3622    The output is an integer that contains the number of bytes used for internal
3623    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3624    
3625      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3626    
3627    The output is an integer that contains the threshold above which the POSIX
3628    interface uses malloc() for output vectors. See item 31 above.
3629    
3630      PCRE_CONFIG_MATCH_LIMIT
3631    
3632    The output is an unsigned integer that contains the default limit of the number
3633    of match() calls in a pcre_exec() execution. See 42 above.
3634    
3635    44. pcretest has been upgraded by the addition of the -C option. This causes it
3636    to extract all the available output from the new pcre_config() function, and to
3637    output it. The program then exits immediately.
3638    
3639    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3640    order to support additional features. One way would have been to define
3641    pcre_exec2() (for example) with extra arguments, but this would not have been
3642    extensible, and would also have required all calls to the original function to
3643    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3644    is used for passing in "extra" data from pcre_study().
3645    
3646    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3647    contains the following fields:
3648    
3649      flags         a bitmap indicating which of the following fields are set
3650      study_data    opaque data from pcre_study()
3651      match_limit   a way of specifying a limit on match() calls for a specific
3652                      call to pcre_exec()
3653      callout_data  data for callouts (see 49 below)
3654    
3655    The flag bits are also defined in pcre.h, and are
3656    
3657      PCRE_EXTRA_STUDY_DATA
3658      PCRE_EXTRA_MATCH_LIMIT
3659      PCRE_EXTRA_CALLOUT_DATA
3660    
3661    The pcre_study() function now returns one of these new pcre_extra blocks, with
3662    the actual study data pointed to by the study_data field, and the
3663    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3664    before. That is, this change is entirely upwards-compatible and requires no
3665    change to existing code.
3666    
3667    If you want to pass in additional data to pcre_exec(), you can either place it
3668    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3669    block.
3670    
3671    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3672    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3673    times with different match limits, until it finds the minimum value needed for
3674    pcre_exec() to complete. The value is then output. This can be instructive; for
3675    most simple matches the number is quite small, but for pathological cases it
3676    gets very large very quickly.
3677    
3678    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3679    returns the size of the data block pointed to by the study_data field in a
3680    pcre_extra block, that is, the value that was passed as the argument to
3681    pcre_malloc() when PCRE was getting memory in which to place the information
3682    created by pcre_study(). The fourth argument should point to a size_t variable.
3683    pcretest has been extended so that this information is shown after a successful
3684    pcre_study() call when information about the compiled regex is being displayed.
3685    
3686    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3687    because what follows is always an absolute path. (Later: it turns out that this
3688    is more than cosmetic for MinGW, because it doesn't like empty path
3689    components.)
3690    
3691    49. Some changes have been made to the callout feature (see 28 above):
3692    
3693    (i)  A callout function now has three choices for what it returns:
3694    
3695           0  =>  success, carry on matching
3696         > 0  =>  failure at this point, but backtrack if possible
3697         < 0  =>  serious error, return this value from pcre_exec()
3698    
3699         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3700         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3701         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3702         use by callout functions. It will never be used by PCRE itself.
3703    
3704    (ii) The pcre_extra structure (see 45 above) has a void * field called
3705         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3706         pcre_callout_block structure has a field of the same name. The contents of
3707         the field passed in the pcre_extra structure are passed to the callout
3708         function in the corresponding field in the callout block. This makes it
3709         easier to use the same callout-containing regex from multiple threads. For
3710         testing, the pcretest program has a new data escape
3711    
3712           \C*n        pass the number n (may be negative) as callout_data
3713    
3714         If the callout function in pcretest receives a non-zero value as
3715         callout_data, it returns that value.
3716    
3717    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3718    there were some redundant $(CFLAGS) in commands that are now specified as
3719    $(LINK), which already includes $(CFLAGS).
3720    
3721    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3722    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3723    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3724    one-byte characters throughout. Note that case-insensitive matching applies
3725    only to characters whose values are less than 256. PCRE doesn't support the
3726    notion of cases for higher-valued characters.
3727    
3728    (i)   A character class whose characters are all within 0-255 is handled as
3729          a bit map, and the map is inverted for negative classes. Previously, a
3730          character > 255 always failed to match such a class; however it should
3731          match if the class was a negative one (e.g. [^ab]). This has been fixed.
3732    
3733    (ii)  A negated character class with a single character < 255 is coded as
3734          "not this character" (OP_NOT). This wasn't working properly when the test
3735          character was multibyte, either singly or repeated.
3736    
3737    (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3738          mode, for example: \x{100}{2,3}.
3739    
3740    (iv)  The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3741          singly or repeated) now correctly test multibyte characters. However,
3742          PCRE doesn't recognize any characters with values greater than 255 as
3743          digits, spaces, or word characters. Such characters always match \D, \S,
3744          and \W, and never match \d, \s, or \w.
3745    
3746    (v)   Classes may now contain characters and character ranges with values
3747          greater than 255. For example: [ab\x{100}-\x{400}].
3748    
3749    (vi)  pcregrep now has a --utf-8 option (synonym -u) which makes it call
3750          PCRE in UTF-8 mode.
3751    
3752    52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3753    PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3754    retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3755    value.)
3756    
3757    53. The single man page has become too large. I have therefore split it up into
3758    a number of separate man pages. These also give rise to individual HTML pages;
3759    these are now put in a separate directory, and there is an index.html page that
3760    lists them all. Some hyperlinking between the pages has been installed.
3761    
3762    54. Added convenience functions for handling named capturing parentheses.
3763    
3764    55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3765    aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3766    true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3767    are faulted.
3768    
3769    56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3770    calling configure. These values are used when compiling the dftables.c program
3771    which is run to generate the source of the default character tables. They
3772    default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3773    you will need to set these values.
3774    
3775    57. Updated the building process for Windows DLL, as provided by Fred Cox.
3776    
3777    
3778    Version 3.9 02-Jan-02
3779    ---------------------
3780    
3781    1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3782    
3783    2. If --disable-static was given, the building process failed when trying to
3784    build pcretest and pcregrep. (For some reason it was using libtool to compile
3785    them, which is not right, as they aren't part of the library.)
3786    
3787    
3788    Version 3.8 18-Dec-01
3789    ---------------------
3790    
3791    1. The experimental UTF-8 code was completely screwed up. It was packing the
3792    bytes in the wrong order. How dumb can you get?
3793    
3794    
3795    Version 3.7 29-Oct-01
3796    ---------------------
3797    
3798    1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3799    This caused pcretest, when used on the test data, to segfault. Unfortunately,
3800    this didn't happen under Solaris 8, where I normally test things.
3801    
3802    2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3803    doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3804    isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3805    this fix an hour or so after the initial 3.7 release.)
3806    
3807    
3808    Version 3.6 23-Oct-01
3809    ---------------------
3810    
3811    1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if