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

Diff of /code/trunk/ChangeLog

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

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