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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.53  
changed lines
  Added in v.691

  ViewVC Help
Powered by ViewVC 1.1.5