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

Diff of /code/trunk/ChangeLog

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

revision 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 663 by ph10, Mon Aug 22 14:53:12 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.20
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.
11    
12    2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no captured
13       substring, but while checking the failing first alternative, substring 1 is
14       temporarily captured. If the output vector supplied to pcre_exec() was not
15       big enough for this capture, the yield of the function was still zero
16       ("insufficient space for captured substrings"). This cannot be totally fixed
17       without adding another stack variable, which seems a lot of expense for a
18       edge case. However, I have improved the situation in cases such as
19       /(a)(b)x|abc/ matched against "abc", where the return code indicates that
20       fewer than the maximum number of slots in the ovector have been set.
21    
22    3. Related to (2) above: when there are more back references in a pattern than
23       slots in the output vector, pcre_exec() uses temporary memory during
24       matching, and copies in the captures as far as possible afterwards. It was
25       using the entire output vector, but this conflicts with the specification
26       that only 2/3 is used for passing back captured substrings. Now it uses only
27       the first 2/3, for compatibility. This is, of course, another edge case.
28    
29    
30    Version 8.13 16-Aug-2011
31    ------------------------
32    
33    1.  The Unicode data tables have been updated to Unicode 6.0.0.
34    
35    2.  Two minor typos in pcre_internal.h have been fixed.
36    
37    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
38        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
39        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
40    
41    4.  There were a number of related bugs in the code for matching backrefences
42        caselessly in UTF-8 mode when codes for the characters concerned were
43        different numbers of bytes. For example, U+023A and U+2C65 are an upper
44        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
45        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
46        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
47        2-byte code at the end of the subject (it thought there wasn't enough data
48        left).
49    
50    5.  Comprehensive information about what went wrong is now returned by
51        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
52        as the output vector has at least 2 elements. The offset of the start of
53        the failing character and a reason code are placed in the vector.
54    
55    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
56        now returned is for the first byte of the failing character, instead of the
57        last byte inspected. This is an incompatible change, but I hope it is small
58        enough not to be a problem. It makes the returned offset consistent with
59        pcre_exec() and pcre_dfa_exec().
60    
61    7.  pcretest now gives a text phrase as well as the error number when
62        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
63        failure, the offset and reason code are output.
64    
65    8.  When \R was used with a maximizing quantifier it failed to skip backwards
66        over a \r\n pair if the subsequent match failed. Instead, it just skipped
67        back over a single character (\n). This seems wrong (because it treated the
68        two characters as a single entity when going forwards), conflicts with the
69        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
70        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
71        has been changed.
72    
73    9.  Some internal refactoring has changed the processing so that the handling
74        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
75        time (the PCRE_DOTALL option was changed this way some time ago: version
76        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
77        which was always a bit of a fudge. It also means that there is one less
78        argument for the match() function, which reduces its stack requirements
79        slightly. This change also fixes an incompatibility with Perl: the pattern
80        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
81    
82    10. More internal refactoring has drastically reduced the number of recursive
83        calls to match() for possessively repeated groups such as (abc)++ when
84        using pcre_exec().
85    
86    11. While implementing 10, a number of bugs in the handling of groups were
87        discovered and fixed:
88    
89        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
90        (a|)*(?1) gave a compile-time internal error.
91        ((a|)+)+  did not notice that the outer group could match an empty string.
92        (^a|^)+   was not marked as anchored.
93        (.*a|.*)+ was not marked as matching at start or after a newline.
94    
95    12. Yet more internal refactoring has removed another argument from the match()
96        function. Special calls to this function are now indicated by setting a
97        value in a variable in the "match data" data block.
98    
99    13. Be more explicit in pcre_study() instead of relying on "default" for
100        opcodes that mean there is no starting character; this means that when new
101        ones are added and accidentally left out of pcre_study(), testing should
102        pick them up.
103    
104    14. The -s option of pcretest has been documented for ages as being an old
105        synonym of -m (show memory usage). I have changed it to mean "force study
106        for every regex", that is, assume /S for every regex. This is similar to -i
107        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
108        using it. It makes it easier to run collections of tests with and without
109        study enabled, and thereby test pcre_study() more easily. All the standard
110        tests are now run with and without -s (but some patterns can be marked as
111        "never study" - see 20 below).
112    
113    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
114        restoration of the capturing data to the outer values was not happening
115        correctly.
116    
117    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
118        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
119        pattern had matched an empty string, and so incorrectly returned a no
120        match.
121    
122    17. There was optimizing code for the last branch of non-capturing parentheses,
123        and also for the obeyed branch of a conditional subexpression, which used
124        tail recursion to cut down on stack usage. Unfortunately, now that there is
125        the possibility of (*THEN) occurring in these branches, tail recursion is
126        no longer possible because the return has to be checked for (*THEN). These
127        two optimizations have therefore been removed.
128    
129    18. If a pattern containing \R was studied, it was assumed that \R always
130        matched two bytes, thus causing the minimum subject length to be
131        incorrectly computed because \R can also match just one byte.
132    
133    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
134        was incorrectly computed.
135    
136    20. If /S is present twice on a test pattern in pcretest input, it now
137        *disables* studying, thereby overriding the use of -s on the command line
138        (see 14 above). This is necessary for one or two tests to keep the output
139        identical in both cases.
140    
141    21. When (*ACCEPT) was used in an assertion that matched an empty string and
142        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
143    
144    22. When an atomic group that contained a capturing parenthesis was
145        successfully matched, but the branch in which it appeared failed, the
146        capturing was not being forgotten if a higher numbered group was later
147        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
148        group 1 to "a", when in fact it should be unset. This applied to multi-
149        branched capturing and non-capturing groups, repeated or not, and also to
150        positive assertions (capturing in negative assertions does not happen
151        in PCRE) and also to nested atomic groups.
152    
153    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
154        subject after a captured substring, to make it easier to tell which of a
155        number of identical substrings has been captured.
156    
157    24. The way atomic groups are processed by pcre_exec() has been changed so that
158        if they are repeated, backtracking one repetition now resets captured
159        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
160        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
161        "aaa". Previously, it would have been "a". As part of this code
162        refactoring, the way recursive calls are handled has also been changed.
163    
164    25. If an assertion condition captured any substrings, they were not passed
165        back unless some other capturing happened later. For example, if
166        (?(?=(a))a) was matched against "a", no capturing was returned.
167    
168    26. When studying a pattern that contained subroutine calls or assertions,
169        the code for finding the minimum length of a possible match was handling
170        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
171        group 1 called group 2 while simultaneously a separate group 2 called group
172        1). A stack overflow occurred in this case. I have fixed this by limiting
173        the recursion depth to 10.
174    
175    27. Updated RunTest.bat in the distribution to the version supplied by Tom
176        Fortmann. This supports explicit test numbers on the command line, and has
177        argument validation and error reporting.
178    
179    28. An instance of \X with an unlimited repeat could fail if at any point the
180        first character it looked at was a mark character.
181    
182    29. Some minor code refactoring concerning Unicode properties and scripts
183        should reduce the stack requirement of match() slightly.
184    
185    30. Added the '=' option to pcretest to check the setting of unused capturing
186        slots at the end of the pattern, which are documented as being -1, but are
187        not included in the return count.
188    
189    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
190        compiled something random. Now it gives a compile-time error (as does
191        Perl).
192    
193    32. A *MARK encountered during the processing of a positive assertion is now
194        recorded and passed back (compatible with Perl).
195    
196    33. If --only-matching or --colour was set on a pcregrep call whose pattern
197        had alternative anchored branches, the search for a second match in a line
198        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
199        matched the line "0102" twice. The same bug affected patterns that started
200        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
201        twice.
202    
203    34. Previously, PCRE did not allow quantification of assertions. However, Perl
204        does, and because of capturing effects, quantifying parenthesized
205        assertions may at times be useful. Quantifiers are now allowed for
206        parenthesized assertions.
207    
208    35. A minor code tidy in pcre_compile() when checking options for \R usage.
209    
210    36. \g was being checked for fancy things in a character class, when it should
211        just be a literal "g".
212    
213    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
214        appearance of a nested POSIX class supersedes an apparent external class.
215        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
216        unescaped square brackets may also appear as part of class names. For
217        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
218        more like Perl. (But see 8.20/1 above.)
219    
220    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
221        was because it thought it was \N{name}, which is not supported).
222    
223    39. Add minix to OS list not supporting the -S option in pcretest.
224    
225    40. PCRE tries to detect cases of infinite recursion at compile time, but it
226        cannot analyze patterns in sufficient detail to catch mutual recursions
227        such as ((?1))((?2)). There is now a runtime test that gives an error if a
228        subgroup is called recursively as a subpattern for a second time at the
229        same position in the subject string. In previous releases this might have
230        been caught by the recursion limit, or it might have run out of stack.
231    
232    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
233        happen only once. PCRE was, however incorrectly giving a compile time error
234        "recursive call could loop indefinitely" because it cannot analyze the
235        pattern in sufficient detail. The compile time test no longer happens when
236        PCRE is compiling a conditional subpattern, but actual runaway loops are
237        now caught at runtime (see 40 above).
238    
239    42. It seems that Perl allows any characters other than a closing parenthesis
240        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
241        has been changed to be the same.
242    
243    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
244        as not to get warnings when autogen.sh is called. Also changed
245        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
246    
247    44. To help people who use pcregrep to scan files containing exceedingly long
248        lines, the following changes have been made:
249    
250        (a) The default value of the buffer size parameter has been increased from
251            8K to 20K. (The actual buffer used is three times this size.)
252    
253        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
254            PCRE is built.
255    
256        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
257            to be set at run time.
258    
259        (d) Numerical values in pcregrep options can be followed by K or M, for
260            example --buffer-size=50K.
261    
262        (e) If a line being scanned overflows pcregrep's buffer, an error is now
263            given and the return code is set to 2.
264    
265    45. Add a pointer to the latest mark to the callout data block.
266    
267    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
268        partial match of an empty string instead of no match. This was specific to
269        the use of ".".
270    
271    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
272        complete match instead of a partial match. This bug was dependent on both
273        the PCRE_UTF8 and PCRE_DOTALL options being set.
274    
275    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
276        starting byte set, because \b was not being ignored.
277    
278    
279    Version 8.12 15-Jan-2011
280    ------------------------
281    
282    1.  Fixed some typos in the markup of the man pages, and wrote a script that
283        checks for such things as part of the documentation building process.
284    
285    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
286        --match-limit and --recursion-limit options (added for 8.11). In
287        particular, this made one of the standard tests fail. (The integer value
288        went into the wrong half of a long int.)
289    
290    3.  If the --colour option was given to pcregrep with -v (invert match), it
291        did strange things, either producing crazy output, or crashing. It should,
292        of course, ignore a request for colour when reporting lines that do not
293        match.
294    
295    4.  Another pcregrep bug caused similar problems if --colour was specified with
296        -M (multiline) and the pattern match finished with a line ending.
297    
298    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
299        matched in multiline mode, the following line was shown as part of the
300        match. This seems wrong, so I have changed it.
301    
302    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
303        the check for further matches in the same line (so they could be coloured)
304        to overrun the end of the current line. If another match was found, it was
305        incorrectly shown (and then shown again when found in the next line).
306    
307    7.  If pcregrep was compiled under Windows, there was a reference to the
308        function pcregrep_exit() before it was defined. I am assuming this was
309        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
310        reported by a user. I've moved the definition above the reference.
311    
312    
313    Version 8.11 10-Dec-2010
314    ------------------------
315    
316    1.  (*THEN) was not working properly if there were untried alternatives prior
317        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
318        backtracked to try for "b" instead of moving to the next alternative branch
319        at the same level (in this case, to look for "c"). The Perl documentation
320        is clear that when (*THEN) is backtracked onto, it goes to the "next
321        alternative in the innermost enclosing group".
322    
323    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
324        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
325        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
326        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
327        (*THEN).
328    
329    3.  If \s appeared in a character class, it removed the VT character from
330        the class, even if it had been included by some previous item, for example
331        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
332        of \s, but is part of the POSIX "space" class.)
333    
334    4.  A partial match never returns an empty string (because you can always
335        match an empty string at the end of the subject); however the checking for
336        an empty string was starting at the "start of match" point. This has been
337        changed to the "earliest inspected character" point, because the returned
338        data for a partial match starts at this character. This means that, for
339        example, /(?<=abc)def/ gives a partial match for the subject "abc"
340        (previously it gave "no match").
341    
342    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
343        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
344        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
345        has an implication that the given string is incomplete (because a partial
346        match is preferred over a full match). For this reason, these items now
347        give a partial match in this situation. [Aside: previously, the one case
348        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
349        match rather than a full match, which was wrong by the old rules, but is
350        now correct.]
351    
352    6.  There was a bug in the handling of #-introduced comments, recognized when
353        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
354        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
355        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
356        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
357        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
358        places in pcre_compile().
359    
360    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
361        comments when looking ahead for named forward references to subpatterns,
362        the only newline sequence it recognized was NL. It now handles newlines
363        according to the set newline convention.
364    
365    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
366        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
367        cater for a lack of strerror(). These oversights have been fixed.
368    
369    9.  Added --match-limit and --recursion-limit to pcregrep.
370    
371    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
372    
373    11. When the -o option was used, pcregrep was setting a return code of 1, even
374        when matches were found, and --line-buffered was not being honoured.
375    
376    12. Added an optional parentheses number to the -o and --only-matching options
377        of pcregrep.
378    
379    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
380        can match an empty string. The code to do it in pcretest and pcredemo
381        needed fixing:
382    
383        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
384            only one byte after an empty string match just before CRLF (this case
385            just got forgotten; "any" and "anycrlf" were OK).
386    
387        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
388            mode when an empty string match preceded an ASCII character followed by
389            a non-ASCII character. (The code for advancing by one character rather
390            than one byte was nonsense.)
391    
392        (c) The pcredemo.c sample program did not have any code at all to handle
393            the cases when CRLF is a valid newline sequence.
394    
395    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
396        as a starting offset was within the subject string. There is now a new
397        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
398        negative or greater than the length of the string. In order to test this,
399        pcretest is extended to allow the setting of negative starting offsets.
400    
401    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
402        starting offset points to the beginning of a UTF-8 character was
403        unnecessarily clumsy. I tidied it up.
404    
405    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
406        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
407    
408    17. Nobody had reported that the --include_dir option, which was added in
409        release 7.7 should have been called --include-dir (hyphen, not underscore)
410        for compatibility with GNU grep. I have changed it to --include-dir, but
411        left --include_dir as an undocumented synonym, and the same for
412        --exclude-dir, though that is not available in GNU grep, at least as of
413        release 2.5.4.
414    
415    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
416        characters from a string of bytes) have been redefined so as not to use
417        loops, in order to improve performance in some environments. At the same
418        time, I abstracted some of the common code into auxiliary macros to save
419        repetition (this should not affect the compiled code).
420    
421    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
422        compile-time error is now given if \c is not followed by an ASCII
423        character, that is, a byte less than 128. (In EBCDIC mode, the code is
424        different, and any byte value is allowed.)
425    
426    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
427        START_OPTIMIZE option, which is now allowed at compile time - but just
428        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
429        to pcregrep and other applications that have no direct access to PCRE
430        options. The new /Y option in pcretest sets this option when calling
431        pcre_compile().
432    
433    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
434        back references. Groups containing recursive back references were forced to
435        be atomic by that change, but in the case of named groups, the amount of
436        memory required was incorrectly computed, leading to "Failed: internal
437        error: code overflow". This has been fixed.
438    
439    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
440        pcretest.c, to avoid build problems in some Borland environments.
441    
442    
443    Version 8.10 25-Jun-2010
444    ------------------------
445    
446    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
447        THEN.
448    
449    2.  (*ACCEPT) was not working when inside an atomic group.
450    
451    3.  Inside a character class, \B is treated as a literal by default, but
452        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
453        causes the error). The code is unchanged, but I tidied the documentation.
454    
455    4.  Inside a character class, PCRE always treated \R and \X as literals,
456        whereas Perl faults them if its -w option is set. I have changed PCRE so
457        that it faults them when PCRE_EXTRA is set.
458    
459    5.  Added support for \N, which always matches any character other than
460        newline. (It is the same as "." when PCRE_DOTALL is not set.)
461    
462    6.  When compiling pcregrep with newer versions of gcc which may have
463        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
464        declared with attribute warn_unused_result" were given. Just casting the
465        result to (void) does not stop the warnings; a more elaborate fudge is
466        needed. I've used a macro to implement this.
467    
468    7.  Minor change to pcretest.c to avoid a compiler warning.
469    
470    8.  Added four artifical Unicode properties to help with an option to make
471        \s etc use properties (see next item). The new properties are: Xan
472        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
473    
474    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
475        use Unicode properties. (*UCP) at the start of a pattern can be used to set
476        this option. Modified pcretest to add /W to test this facility. Added
477        REG_UCP to make it available via the POSIX interface.
478    
479    10. Added --line-buffered to pcregrep.
480    
481    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
482        studied, and the match started with a letter with a code point greater than
483        127 whose first byte was different to the first byte of the other case of
484        the letter, the other case of this starting letter was not recognized
485        (#976).
486    
487    12. If a pattern that was studied started with a repeated Unicode property
488        test, for example, \p{Nd}+, there was the theoretical possibility of
489        setting up an incorrect bitmap of starting bytes, but fortunately it could
490        not have actually happened in practice until change 8 above was made (it
491        added property types that matched character-matching opcodes).
492    
493    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
494        possible starting bytes for non-anchored patterns.
495    
496    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
497        \R, and also a number of cases that involve Unicode properties, both
498        explicit and implicit when PCRE_UCP is set.
499    
500    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
501        input, it could crash or give wrong results if characters with values
502        greater than 0xc0 were present in the subject string. (Detail: it assumed
503        UTF-8 input when processing these items.)
504    
505    16. Added a lot of (int) casts to avoid compiler warnings in systems where
506        size_t is 64-bit (#991).
507    
508    17. Added a check for running out of memory when PCRE is compiled with
509        --disable-stack-for-recursion (#990).
510    
511    18. If the last data line in a file for pcretest does not have a newline on
512        the end, a newline was missing in the output.
513    
514    19. The default pcre_chartables.c file recognizes only ASCII characters (values
515        less than 128) in its various bitmaps. However, there is a facility for
516        generating tables according to the current locale when PCRE is compiled. It
517        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
518        characters, are recognized by isspace() and therefore were getting set in
519        these tables, and indeed these tables seem to approximate to ISO 8859. This
520        caused a problem in UTF-8 mode when pcre_study() was used to create a list
521        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
522        which of course cannot start UTF-8 characters. I have changed the code so
523        that only real ASCII characters (less than 128) and the correct starting
524        bytes for UTF-8 encodings are set for characters greater than 127 when in
525        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
526        altogether.)
527    
528    20. Added the /T option to pcretest so as to be able to run tests with non-
529        standard character tables, thus making it possible to include the tests
530        used for 19 above in the standard set of tests.
531    
532    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
533        reference to a subpattern the other side of a comment that contains an
534        opening parenthesis caused either an internal compiling error, or a
535        reference to the wrong subpattern.
536    
537    
538    Version 8.02 19-Mar-2010
539    ------------------------
540    
541    1.  The Unicode data tables have been updated to Unicode 5.2.0.
542    
543    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
544        configured.
545    
546    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
547        original author of that file, following a query about its status.
548    
549    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
550        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
551    
552    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
553        quantifier applied to a forward-referencing subroutine call, could compile
554        incorrect code or give the error "internal error: previously-checked
555        referenced subpattern not found".
556    
557    6.  Both MS Visual Studio and Symbian OS have problems with initializing
558        variables to point to external functions. For these systems, therefore,
559        pcre_malloc etc. are now initialized to local functions that call the
560        relevant global functions.
561    
562    7.  There were two entries missing in the vectors called coptable and poptable
563        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
564        I've fixed the data, and added a kludgy way of testing at compile time that
565        the lengths are correct (equal to the number of opcodes).
566    
567    8.  Following on from 7, I added a similar kludge to check the length of the
568        eint vector in pcreposix.c.
569    
570    9.  Error texts for pcre_compile() are held as one long string to avoid too
571        much relocation at load time. To find a text, the string is searched,
572        counting zeros. There was no check for running off the end of the string,
573        which could happen if a new error number was added without updating the
574        string.
575    
576    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
577    
578    11. \K was not working if it appeared in an atomic group or in a group that
579        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
580        \K is "not well defined" if used in an assertion. PCRE now accepts it if
581        the assertion is positive, but not if it is negative.
582    
583    12. Change 11 fortuitously reduced the size of the stack frame used in the
584        "match()" function of pcre_exec.c by one pointer. Forthcoming
585        implementation of support for (*MARK) will need an extra pointer on the
586        stack; I have reserved it now, so that the stack frame size does not
587        decrease.
588    
589    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
590        item in branch that calls a recursion is a subroutine call - as in the
591        second branch in the above example - was incorrectly given the compile-
592        time error "recursive call could loop indefinitely" because pcre_compile()
593        was not correctly checking the subroutine for matching a non-empty string.
594    
595    14. The checks for overrunning compiling workspace could trigger after an
596        overrun had occurred. This is a "should never occur" error, but it can be
597        triggered by pathological patterns such as hundreds of nested parentheses.
598        The checks now trigger 100 bytes before the end of the workspace.
599    
600    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
601    
602    
603    Version 8.01 19-Jan-2010
604    ------------------------
605    
606    1.  If a pattern contained a conditional subpattern with only one branch (in
607        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
608        computed the wrong minimum data length (which is of course zero for such
609        subpatterns). This could cause incorrect "no match" results.
610    
611    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
612        the pattern is reset in the first branch, pcre_compile() failed with
613        "internal error: code overflow at offset...". This happened only when
614        the reset was to the original external option setting. (An optimization
615        abstracts leading options settings into an external setting, which was the
616        cause of this.)
617    
618    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
619        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
620        assertion pattern did not match (meaning that the assertion was true), it
621        was incorrectly treated as false if the SKIP had been reached during the
622        matching. This also applied to assertions used as conditions.
623    
624    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
625        assertion subpattern, including such a pattern used as a condition,
626        unpredictable results occurred, instead of the error return
627        PCRE_ERROR_DFA_UITEM.
628    
629    5.  The C++ GlobalReplace function was not working like Perl for the special
630        situation when an empty string is matched. It now does the fancy magic
631        stuff that is necessary.
632    
633    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
634        removed. (These were left over from very, very early versions of PCRE.)
635    
636    7.  Some cosmetic changes to the code to make life easier when compiling it
637        as part of something else:
638    
639        (a) Change DEBUG to PCRE_DEBUG.
640    
641        (b) In pcre_compile(), rename the member of the "branch_chain" structure
642            called "current" as "current_branch", to prevent a collision with the
643            Linux macro when compiled as a kernel module.
644    
645        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
646            prevent a collision with the Linux macro when compiled as a kernel
647            module.
648    
649    8.  In pcre_compile() there are some checks for integer overflows that used to
650        cast potentially large values to (double). This has been changed to that
651        when building, a check for int64_t is made, and if it is found, it is used
652        instead, thus avoiding the use of floating point arithmetic. (There is no
653        other use of FP in PCRE.) If int64_t is not found, the fallback is to
654        double.
655    
656    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
657        2005 (difference between two addresses compared to an unsigned value).
658    
659    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
660        custom one, because of the following reported problem in Windows:
661    
662          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
663              under Win32.
664          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
665              therefore missing the function definition.
666          - The compiler thus generates a "C" signature for the test function.
667          - The linker fails to find the "C" function.
668          - PCRE fails to configure if asked to do so against libbz2.
669    
670    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
671        messages were output:
672    
673          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
674          rerunning libtoolize, to keep the correct libtool macros in-tree.
675          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
676    
677        I have done both of these things.
678    
679    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
680        most of the time, it *can* run out if it is given a pattern that contains a
681        runaway infinite recursion. I updated the discussion in the pcrestack man
682        page.
683    
684    13. Now that we have gone to the x.xx style of version numbers, the minor
685        version may start with zero. Using 08 or 09 is a bad idea because users
686        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
687        interpreted as invalid octal numbers. I've updated the previous comment in
688        configure.ac, and also added a check that gives an error if 08 or 09 are
689        used.
690    
691    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
692        causing partial matching to fail when the end of the subject matched \W
693        in a UTF-8 pattern where \W was quantified with a minimum of 3.
694    
695    15. There were some discrepancies between the declarations in pcre_internal.h
696        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
697        their definitions. The declarations used "const uschar *" and the
698        definitions used USPTR. Even though USPTR is normally defined as "const
699        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
700        reported that: "This difference in casting confuses some C++ compilers, for
701        example, SunCC recognizes above declarations as different functions and
702        generates broken code for hbpcre." I have changed the declarations to use
703        USPTR.
704    
705    16. GNU libtool is named differently on some systems. The autogen.sh script now
706        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
707        (FreeBSD).
708    
709    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
710        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
711        comment: "Figure out how to create a longlong from a string: strtoll and
712        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
713        instance, but it only takes 2 args instead of 3!"
714    
715    18. A subtle bug concerned with back references has been fixed by a change of
716        specification, with a corresponding code fix. A pattern such as
717        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
718        refers, was giving matches when it shouldn't. For example, xa=xaaa would
719        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
720        same bug. Such groups have to be quantified to be useful, or contained
721        inside another quantified group. (If there's no repetition, the reference
722        can never match.) The problem arises because, having left the group and
723        moved on to the rest of the pattern, a later failure that backtracks into
724        the group uses the captured value from the final iteration of the group
725        rather than the correct earlier one. I have fixed this in PCRE by forcing
726        any group that contains a reference to itself to be an atomic group; that
727        is, there cannot be any backtracking into it once it has completed. This is
728        similar to recursive and subroutine calls.
729    
730    
731    Version 8.00 19-Oct-09
732    ----------------------
733    
734    1.  The table for translating pcre_compile() error codes into POSIX error codes
735        was out-of-date, and there was no check on the pcre_compile() error code
736        being within the table. This could lead to an OK return being given in
737        error.
738    
739    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
740        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
741        in a Windows environment.
742    
743    3.  The pcregrep --count option prints the count for each file even when it is
744        zero, as does GNU grep. However, pcregrep was also printing all files when
745        --files-with-matches was added. Now, when both options are given, it prints
746        counts only for those files that have at least one match. (GNU grep just
747        prints the file name in this circumstance, but including the count seems
748        more useful - otherwise, why use --count?) Also ensured that the
749        combination -clh just lists non-zero counts, with no names.
750    
751    4.  The long form of the pcregrep -F option was incorrectly implemented as
752        --fixed_strings instead of --fixed-strings. This is an incompatible change,
753        but it seems right to fix it, and I didn't think it was worth preserving
754        the old behaviour.
755    
756    5.  The command line items --regex=pattern and --regexp=pattern were not
757        recognized by pcregrep, which required --regex pattern or --regexp pattern
758        (with a space rather than an '='). The man page documented the '=' forms,
759        which are compatible with GNU grep; these now work.
760    
761    6.  No libpcreposix.pc file was created for pkg-config; there was just
762        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
763    
764    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
765        when UCP support is not needed, by modifying the Python script that
766        generates it from Unicode data files. This should not matter if the module
767        is correctly used as a library, but I received one complaint about 50K of
768        unwanted data. My guess is that the person linked everything into his
769        program rather than using a library. Anyway, it does no harm.
770    
771    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
772        was a minimum greater than 1 for a wide character in a possessive
773        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
774        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
775        character. Chaos in the form of incorrect output or a compiling loop could
776        result.
777    
778    9.  The restrictions on what a pattern can contain when partial matching is
779        requested for pcre_exec() have been removed. All patterns can now be
780        partially matched by this function. In addition, if there are at least two
781        slots in the offset vector, the offset of the earliest inspected character
782        for the match and the offset of the end of the subject are set in them when
783        PCRE_ERROR_PARTIAL is returned.
784    
785    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
786        synonymous with PCRE_PARTIAL, for backwards compatibility, and
787        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
788        and may be more useful for multi-segment matching.
789    
790    11. Partial matching with pcre_exec() is now more intuitive. A partial match
791        used to be given if ever the end of the subject was reached; now it is
792        given only if matching could not proceed because another character was
793        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
794        string "Z", which now yields "no match" instead of "partial match". In the
795        case of pcre_dfa_exec(), "no match" is given if every matching path for the
796        final character ended with (*FAIL).
797    
798    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
799        if the pattern had a "must contain" character that was already found in the
800        earlier partial match, unless partial matching was again requested. For
801        example, with the pattern /dog.(body)?/, the "must contain" character is
802        "g". If the first part-match was for the string "dog", restarting with
803        "sbody" failed. This bug has been fixed.
804    
805    13. The string returned by pcre_dfa_exec() after a partial match has been
806        changed so that it starts at the first inspected character rather than the
807        first character of the match. This makes a difference only if the pattern
808        starts with a lookbehind assertion or \b or \B (\K is not supported by
809        pcre_dfa_exec()). It's an incompatible change, but it makes the two
810        matching functions compatible, and I think it's the right thing to do.
811    
812    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
813        so that the demonstration program is easily available in environments where
814        PCRE has not been installed from source.
815    
816    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
817        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
818        library.
819    
820    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
821        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
822        is not the first non-POSIX option to be added. Clearly some people find
823        these options useful.
824    
825    17. If a caller to the POSIX matching function regexec() passes a non-zero
826        value for nmatch with a NULL value for pmatch, the value of
827        nmatch is forced to zero.
828    
829    18. RunGrepTest did not have a test for the availability of the -u option of
830        the diff command, as RunTest does. It now checks in the same way as
831        RunTest, and also checks for the -b option.
832    
833    19. If an odd number of negated classes containing just a single character
834        interposed, within parentheses, between a forward reference to a named
835        subpattern and the definition of the subpattern, compilation crashed with
836        an internal error, complaining that it could not find the referenced
837        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
838        [The bug was that it was starting one character too far in when skipping
839        over the character class, thus treating the ] as data rather than
840        terminating the class. This meant it could skip too much.]
841    
842    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
843        /g option in pcretest when the pattern contains \K, which makes it possible
844        to have an empty string match not at the start, even when the pattern is
845        anchored. Updated pcretest and pcredemo to use this option.
846    
847    21. If the maximum number of capturing subpatterns in a recursion was greater
848        than the maximum at the outer level, the higher number was returned, but
849        with unset values at the outer level. The correct (outer level) value is
850        now given.
851    
852    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
853        PCRE did not set those parentheses (unlike Perl). I have now found a way to
854        make it do so. The string so far is captured, making this feature
855        compatible with Perl.
856    
857    23. The tests have been re-organized, adding tests 11 and 12, to make it
858        possible to check the Perl 5.10 features against Perl 5.10.
859    
860    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
861        pattern matches a fixed length string. PCRE did not allow this; now it
862        does. Neither allows recursion.
863    
864    25. I finally figured out how to implement a request to provide the minimum
865        length of subject string that was needed in order to match a given pattern.
866        (It was back references and recursion that I had previously got hung up
867        on.) This code has now been added to pcre_study(); it finds a lower bound
868        to the length of subject needed. It is not necessarily the greatest lower
869        bound, but using it to avoid searching strings that are too short does give
870        some useful speed-ups. The value is available to calling programs via
871        pcre_fullinfo().
872    
873    26. While implementing 25, I discovered to my embarrassment that pcretest had
874        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
875        study optimizations had never been tested with that matching function.
876        Oops. What is worse, even when it was passed study data, there was a bug in
877        pcre_dfa_exec() that meant it never actually used it. Double oops. There
878        were also very few tests of studied patterns with pcre_dfa_exec().
879    
880    27. If (?| is used to create subpatterns with duplicate numbers, they are now
881        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
882        on the other side of the coin, they are no longer allowed to have different
883        names, because these cannot be distinguished in PCRE, and this has caused
884        confusion. (This is a difference from Perl.)
885    
886    28. When duplicate subpattern names are present (necessarily with different
887        numbers, as required by 27 above), and a test is made by name in a
888        conditional pattern, either for a subpattern having been matched, or for
889        recursion in such a pattern, all the associated numbered subpatterns are
890        tested, and the overall condition is true if the condition is true for any
891        one of them. This is the way Perl works, and is also more like the way
892        testing by number works.
893    
894    
895    Version 7.9 11-Apr-09
896    ---------------------
897    
898    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
899        (pcretest), all targets were linked against these libraries. This included
900        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
901        libraries. This caused unwanted dependencies to be created. This problem
902        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
903        pcretest is linked with readline.
904    
905    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
906        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
907        moved outside it again, because FALSE and TRUE are already defined in AIX,
908        but BOOL is not.
909    
910    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
911        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
912    
913    4.  The pcregrep documentation said spaces were inserted as well as colons (or
914        hyphens) following file names and line numbers when outputting matching
915        lines. This is not true; no spaces are inserted. I have also clarified the
916        wording for the --colour (or --color) option.
917    
918    5.  In pcregrep, when --colour was used with -o, the list of matching strings
919        was not coloured; this is different to GNU grep, so I have changed it to be
920        the same.
921    
922    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
923        each matching line was coloured. Now it goes on to look for further matches
924        of any of the test patterns, which is the same behaviour as GNU grep.
925    
926    7.  A pattern that could match an empty string could cause pcregrep to loop; it
927        doesn't make sense to accept an empty string match in pcregrep, so I have
928        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
929        seems to be how GNU grep behaves.
930    
931    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
932        start or after a newline", because the conditional assertion was not being
933        correctly handled. The rule now is that both the assertion and what follows
934        in the first alternative must satisfy the test.
935    
936    9.  If auto-callout was enabled in a pattern with a conditional group whose
937        condition was an assertion, PCRE could crash during matching, both with
938        pcre_exec() and pcre_dfa_exec().
939    
940    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
941        used for matching.
942    
943    11. Unicode property support in character classes was not working for
944        characters (bytes) greater than 127 when not in UTF-8 mode.
945    
946    12. Added the -M command line option to pcretest.
947    
948    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
949    
950    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
951    
952    16. Added comments and documentation about mis-use of no_arg in the C++
953        wrapper.
954    
955    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
956        from Martin Jerabek that uses macro names for all relevant character and
957        string constants.
958    
959    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
960        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
961        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
962        these, but not everybody uses configure.
963    
964    19. A conditional group that had only one branch was not being correctly
965        recognized as an item that could match an empty string. This meant that an
966        enclosing group might also not be so recognized, causing infinite looping
967        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
968        with the subject "ab", where knowledge that the repeated group can match
969        nothing is needed in order to break the loop.
970    
971    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
972        exec(), but without supplying a callout function, matching went wrong.
973    
974    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
975        leak if the size of the offset vector was greater than 30. When the vector
976        is smaller, the saved offsets during recursion go onto a local stack
977        vector, but for larger vectors malloc() is used. It was failing to free
978        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
979        error, in fact).
980    
981    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
982        heapframe that is used only when UTF-8 support is enabled. This caused no
983        problem, but was untidy.
984    
985    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
986        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
987        included within another project.
988    
989    24. Steven Van Ingelgem's patches to add more options to the CMake support,
990        slightly modified by me:
991    
992          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
993              not building pcregrep.
994    
995          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
996              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
997    
998    25. Forward references, both numeric and by name, in patterns that made use of
999        duplicate group numbers, could behave incorrectly or give incorrect errors,
1000        because when scanning forward to find the reference group, PCRE was not
1001        taking into account the duplicate group numbers. A pattern such as
1002        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1003    
1004    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1005        the feature of a custom pointer more persuasive (as requested by a user).
1006    
1007    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1008        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1009        pre-defined.
1010    
1011    28. Added support for (*UTF8) at the start of a pattern.
1012    
1013    29. Arrange for flags added by the "release type" setting in CMake to be shown
1014        in the configuration summary.
1015    
1016    
1017    Version 7.8 05-Sep-08
1018    ---------------------
1019    
1020    1.  Replaced UCP searching code with optimized version as implemented for Ad
1021        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1022        stage table and inline lookup instead of a function, giving speed ups of 2
1023        to 5 times on some simple patterns that I tested. Permission was given to
1024        distribute the MultiStage2.py script that generates the tables (it's not in
1025        the tarball, but is in the Subversion repository).
1026    
1027    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1028        scripts.
1029    
1030    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1031        a group with a zero qualifier. The result of the study could be incorrect,
1032        or the function might crash, depending on the pattern.
1033    
1034    4.  Caseless matching was not working for non-ASCII characters in back
1035        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1036        It now works when Unicode Property Support is available.
1037    
1038    5.  In pcretest, an escape such as \x{de} in the data was always generating
1039        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1040        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1041        truncation.
1042    
1043    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1044    
1045    7.  Added two (int) casts to pcregrep when printing the difference of two
1046        pointers, in case they are 64-bit values.
1047    
1048    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1049        test 2 if it fails.
1050    
1051    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1052        and a #define of that name to empty if it is not externally set. This is to
1053        allow users of MSVC to set it if necessary.
1054    
1055    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1056        the convenience functions in the pcre_get.c source file.
1057    
1058    11. An option change at the start of a pattern that had top-level alternatives
1059        could cause overwriting and/or a crash. This command provoked a crash in
1060        some environments:
1061    
1062          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1063    
1064        This potential security problem was recorded as CVE-2008-2371.
1065    
1066    12. For a pattern where the match had to start at the beginning or immediately
1067        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1068        pcre_dfa_exec() could read past the end of the passed subject if there was
1069        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1070        pcretest so that it places the subject at the end of its malloc-ed buffer.
1071    
1072    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1073        exec() might read past the end of the data buffer in UTF-8 mode.
1074    
1075    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1076        the data contained the byte 0x85 as part of a UTF-8 character within its
1077        first line. This applied both to normal and DFA matching.
1078    
1079    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1080        /^[^d]*?$/8 failed to match "abc".
1081    
1082    16. Added a missing copyright notice to pcrecpp_internal.h.
1083    
1084    17. Make it more clear in the documentation that values returned from
1085        pcre_exec() in ovector are byte offsets, not character counts.
1086    
1087    18. Tidied a few places to stop certain compilers from issuing warnings.
1088    
1089    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1090        supplied by Stefan Weber. I made a further small update for 7.8 because
1091        there is a change of source arrangements: the pcre_searchfuncs.c module is
1092        replaced by pcre_ucd.c.
1093    
1094    
1095    Version 7.7 07-May-08
1096    ---------------------
1097    
1098    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1099        a string to a long long, pretend we don't even have a long long." This is
1100        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1101    
1102    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1103        pre-7.6 versions, which defined a global no_arg variable instead of putting
1104        it in the RE class. (See also #8 below.)
1105    
1106    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1107        Lopes.
1108    
1109    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1110    
1111        (1) The include/exclude patterns were being applied to the whole pathnames
1112            of files, instead of just to the final components.
1113    
1114        (2) If there was more than one level of directory, the subdirectories were
1115            skipped unless they satisfied the include/exclude conditions. This is
1116            inconsistent with GNU grep (and could even be seen as contrary to the
1117            pcregrep specification - which I improved to make it absolutely clear).
1118            The action now is always to scan all levels of directory, and just
1119            apply the include/exclude patterns to regular files.
1120    
1121    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1122        --exclude_dir in the tests to avoid scanning .svn directories.
1123    
1124    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1125        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1126        doesn't support NULs in patterns.
1127    
1128    7.  Added some missing "const"s to declarations of static tables in
1129        pcre_compile.c and pcre_dfa_exec.c.
1130    
1131    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1132        caused by fix #2  above. (Subsequently also a second patch to fix the
1133        first patch. And a third patch - this was a messy problem.)
1134    
1135    9.  Applied Craig's patch to remove the use of push_back().
1136    
1137    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1138        matching function regexec().
1139    
1140    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1141        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1142        references. PCRE supports relative numbers with this syntax (I don't think
1143        Oniguruma does).
1144    
1145    12. Previously, a group with a zero repeat such as (...){0} was completely
1146        omitted from the compiled regex. However, this means that if the group
1147        was called as a subroutine from elsewhere in the pattern, things went wrong
1148        (an internal error was given). Such groups are now left in the compiled
1149        pattern, with a new opcode that causes them to be skipped at execution
1150        time.
1151    
1152    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1153        to the way PCRE behaves:
1154    
1155        (a) A lone ] character is dis-allowed (Perl treats it as data).
1156    
1157        (b) A back reference to an unmatched subpattern matches an empty string
1158            (Perl fails the current match path).
1159    
1160        (c) A data ] in a character class must be notated as \] because if the
1161            first data character in a class is ], it defines an empty class. (In
1162            Perl it is not possible to have an empty class.) The empty class []
1163            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1164            The negative empty class [^] matches any one character, independently
1165            of the DOTALL setting.
1166    
1167    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1168        non-existent subpattern following a character class starting with ']' and
1169        containing () gave an internal compiling error instead of "reference to
1170        non-existent subpattern". Fortunately, when the pattern did exist, the
1171        compiled code was correct. (When scanning forwards to check for the
1172        existencd of the subpattern, it was treating the data ']' as terminating
1173        the class, so got the count wrong. When actually compiling, the reference
1174        was subsequently set up correctly.)
1175    
1176    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1177        it was being rejected as not supported by pcre_dfa_exec(), even though
1178        other assertions are supported. I have made pcre_dfa_exec() support
1179        (*FAIL).
1180    
1181    16. The implementation of 13c above involved the invention of a new opcode,
1182        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1183        cannot be changed at match time, I realized I could make a small
1184        improvement to matching performance by compiling OP_ALLANY instead of
1185        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1186        on the OP_ANY path.
1187    
1188    17. Compiling pcretest on Windows with readline support failed without the
1189        following two fixes: (1) Make the unistd.h include conditional on
1190        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1191    
1192    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1193        ncurses library to be included for pcretest when ReadLine support is
1194        requested, but also to allow for it to be overridden. This patch came from
1195        Daniel Bergström.
1196    
1197    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1198        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1199        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1200        spotting this.
1201    
1202    
1203    Version 7.6 28-Jan-08
1204    ---------------------
1205    
1206    1.  A character class containing a very large number of characters with
1207        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1208        overflow.
1209    
1210    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1211        HAVE_LONG_LONG is not defined.
1212    
1213    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1214        bring them up to date and include new features. This patch includes:
1215    
1216        - Fixed PH's badly added libz and libbz2 support.
1217        - Fixed a problem with static linking.
1218        - Added pcredemo. [But later removed - see 7 below.]
1219        - Fixed dftables problem and added an option.
1220        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1221            HAVE_LONG_LONG.
1222        - Added readline support for pcretest.
1223        - Added an listing of the option settings after cmake has run.
1224    
1225    4.  A user submitted a patch to Makefile that makes it easy to create
1226        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1227        Makefile.am that cause it to include this special target, without
1228        affecting anything else. Note that the same mingw target plus all
1229        the other distribution libraries and programs are now supported
1230        when configuring with CMake (see 6 below) instead of with
1231        Configure/Make.
1232    
1233    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1234        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1235        exported in the Windows port". It has not yet been confirmed that the patch
1236        solves the problem, but it does no harm.
1237    
1238    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1239        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1240        with CMake, and also correct the comment about stack recursion.
1241    
1242    7.  Remove the automatic building of pcredemo from the ./configure system and
1243        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1244        of a program that users should build themselves after PCRE is installed, so
1245        building it automatically is not really right. What is more, it gave
1246        trouble in some build environments.
1247    
1248    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1249    
1250    
1251    Version 7.5 10-Jan-08
1252    ---------------------
1253    
1254    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1255        values in parens when parsing an RE using the C++ wrapper."
1256    
1257    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1258        Characters greater than 255 were excluded from the class instead of being
1259        included.
1260    
1261    3.  The same bug as (2) above applied to negated POSIX classes such as
1262        [:^space:].
1263    
1264    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1265        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1266        I have changed it.
1267    
1268    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1269        first named subpattern) and a construct such as (?&a) would reference the
1270        first named subpattern whose name started with "a" (in other words, the
1271        length check was missing). Both these problems are fixed. "Subpattern name
1272        expected" is now given for (?&) (a zero-length name), and this patch also
1273        makes it give the same error for \k'' (previously it complained that that
1274        was a reference to a non-existent subpattern).
1275    
1276    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1277        this is right because (?- can be followed by option settings as well as by
1278        digits. I have, however, made the messages clearer.
1279    
1280    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1281        than the number used in the conditional) now cause a compile-time error.
1282        This is actually not compatible with Perl, which accepts such patterns, but
1283        treats the conditional as always being FALSE (as PCRE used to), but it
1284        seems to me that giving a diagnostic is better.
1285    
1286    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1287        and messages.
1288    
1289    9.  Fix two occurrences of "backslash" in comments that should have been
1290        "backspace".
1291    
1292    10. Remove two redundant lines of code that can never be obeyed (their function
1293        was moved elsewhere).
1294    
1295    11. The program that makes PCRE's Unicode character property table had a bug
1296        which caused it to generate incorrect table entries for sequences of
1297        characters that have the same character type, but are in different scripts.
1298        It amalgamated them into a single range, with the script of the first of
1299        them. In other words, some characters were in the wrong script. There were
1300        thirteen such cases, affecting characters in the following ranges:
1301    
1302          U+002b0 - U+002c1
1303          U+0060c - U+0060d
1304          U+0061e - U+00612
1305          U+0064b - U+0065e
1306          U+0074d - U+0076d
1307          U+01800 - U+01805
1308          U+01d00 - U+01d77
1309          U+01d9b - U+01dbf
1310          U+0200b - U+0200f
1311          U+030fc - U+030fe
1312          U+03260 - U+0327f
1313          U+0fb46 - U+0fbb1
1314          U+10450 - U+1049d
1315    
1316    12. The -o option (show only the matching part of a line) for pcregrep was not
1317        compatible with GNU grep in that, if there was more than one match in a
1318        line, it showed only the first of them. It now behaves in the same way as
1319        GNU grep.
1320    
1321    13. If the -o and -v options were combined for pcregrep, it printed a blank
1322        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1323        does the same. The return code can be used to tell if there were any
1324        non-matching lines.
1325    
1326    14. Added --file-offsets and --line-offsets to pcregrep.
1327    
1328    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1329        infinitely looping recursion. The bug was that positive lookaheads were not
1330        being skipped when checking for a possible empty match (negative lookaheads
1331        and both kinds of lookbehind were skipped).
1332    
1333    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1334        inclusion of <windows.h> to before rather than after the definition of
1335        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1336    
1337    17. Specifying a possessive quantifier with a specific limit for a Unicode
1338        character property caused pcre_compile() to compile bad code, which led at
1339        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1340        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1341        caused the error; without that there was no problem.
1342    
1343    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1344    
1345    19. Added --enable-pcretest-libreadline.
1346    
1347    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1348        RE::GlobalReplace(). As a result, the number of replacements returned was
1349        double what it should be. I removed one of the increments, but Craig sent a
1350        later patch that removed the other one (the right fix) and added unit tests
1351        that check the return values (which was not done before).
1352    
1353    21. Several CMake things:
1354    
1355        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1356            the names libpcre and libpcreposix, not just pcre and pcreposix.
1357    
1358        (2) The above change means that pcretest and pcregrep are now correctly
1359            linked with the newly-built libraries, not previously installed ones.
1360    
1361        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1362    
1363    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1364        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1365        UTF-8 newline character). The key issue is that the pattern starts .*;
1366        this means that the match must be either at the beginning, or after a
1367        newline. The bug was in the code for advancing after a failed match and
1368        checking that the new position followed a newline. It was not taking
1369        account of UTF-8 characters correctly.
1370    
1371    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1372        character classes. PCRE was not treating the sequence [:...:] as a
1373        character class unless the ... were all letters. Perl, however, seems to
1374        allow any characters between [: and :], though of course it rejects as
1375        unknown any "names" that contain non-letters, because all the known class
1376        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1377        for example, whereas PCRE did not - it did not recognize a POSIX character
1378        class. This seemed a bit dangerous, so the code has been changed to be
1379        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1380        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1381        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1382        Perl does, and where it didn't before.
1383    
1384    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1385        Windows environments %n is disabled by default.
1386    
1387    
1388    Version 7.4 21-Sep-07
1389    ---------------------
1390    
1391    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1392        means that a class such as [\s] counted as "explicit reference to CR or
1393        LF". That isn't really right - the whole point of the change was to try to
1394        help when there was an actual mention of one of the two characters. So now
1395        the change happens only if \r or \n (or a literal CR or LF) character is
1396        encountered.
1397    
1398    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1399        of both had grown to the point where there were only 3 bits left.
1400        Fortunately, there was spare space in the data structure, and so I have
1401        moved the internal flags into a new 16-bit field to free up more option
1402        bits.
1403    
1404    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1405        but did not set the internal JCHANGED flag - either of these is enough to
1406        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1407        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1408        start sets both bits.
1409    
1410    4.  Added options (at build time, compile time, exec time) to change \R from
1411        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1412    
1413    5.  doc/pcresyntax.html was missing from the distribution.
1414    
1415    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1416        compatibility, even though it is no longer used.
1417    
1418    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1419        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1420        windows.h file is present (where different names are used). [This was
1421        reversed later after testing - see 16 below.]
1422    
1423    8.  Changed all #include <config.h> to #include "config.h". There were also
1424        some further <pcre.h> cases that I changed to "pcre.h".
1425    
1426    9.  When pcregrep was used with the --colour option, it missed the line ending
1427        sequence off the lines that it output.
1428    
1429    10. It was pointed out to me that arrays of string pointers cause lots of
1430        relocations when a shared library is dynamically loaded. A technique of
1431        using a single long string with a table of offsets can drastically reduce
1432        these. I have refactored PCRE in four places to do this. The result is
1433        dramatic:
1434    
1435          Originally:                          290
1436          After changing UCP table:            187
1437          After changing error message table:   43
1438          After changing table of "verbs"       36
1439          After changing table of Posix names   22
1440    
1441        Thanks to the folks working on Gregex for glib for this insight.
1442    
1443    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1444        unicode-properties was also set.
1445    
1446    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1447    
1448    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1449        checked only for CRLF.
1450    
1451    14. Added casts to pcretest.c to avoid compiler warnings.
1452    
1453    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1454    
1455    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1456        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1457        entirely. This removes changes made in 7 above.
1458    
1459    17. The CMake files have been updated, and there is now more information about
1460        building with CMake in the NON-UNIX-USE document.
1461    
1462    
1463    Version 7.3 28-Aug-07
1464    ---------------------
1465    
1466     1. In the rejigging of the build system that eventually resulted in 7.1, the
1467        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1468        brackets there is not right, since it causes compilers to look for an
1469        installed pcre.h, not the version that is in the source that is being
1470        compiled (which of course may be different). I have changed it back to:
1471    
1472          #include "pcre.h"
1473    
1474        I have a vague recollection that the change was concerned with compiling in
1475        different directories, but in the new build system, that is taken care of
1476        by the VPATH setting the Makefile.
1477    
1478     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1479        when the subject happened to end in the byte 0x85 (e.g. if the last
1480        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1481        characters but of course it shouldn't be taken as a newline when it is part
1482        of another character. The bug was that, for an unlimited repeat of . in
1483        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1484        characters when looking for a newline.
1485    
1486     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1487    
1488     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1489        in debug output.
1490    
1491     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1492        long printing in the pcrecpp unittest when running under MinGW.
1493    
1494     6. ESC_K was left out of the EBCDIC table.
1495    
1496     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1497        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1498        limit also applies to "virtual nesting" when a pattern is recursive, and in
1499        this case 1000 isn't so big. I have been able to remove this limit at the
1500        expense of backing off one optimization in certain circumstances. Normally,
1501        when pcre_exec() would call its internal match() function recursively and
1502        immediately return the result unconditionally, it uses a "tail recursion"
1503        feature to save stack. However, when a subpattern that can match an empty
1504        string has an unlimited repetition quantifier, it no longer makes this
1505        optimization. That gives it a stack frame in which to save the data for
1506        checking that an empty string has been matched. Previously this was taken
1507        from the 1000-entry workspace that had been reserved. So now there is no
1508        explicit limit, but more stack is used.
1509    
1510     8. Applied Daniel's patches to solve problems with the import/export magic
1511        syntax that is required for Windows, and which was going wrong for the
1512        pcreposix and pcrecpp parts of the library. These were overlooked when this
1513        problem was solved for the main library.
1514    
1515     9. There were some crude static tests to avoid integer overflow when computing
1516        the size of patterns that contain repeated groups with explicit upper
1517        limits. As the maximum quantifier is 65535, the maximum group length was
1518        set at 30,000 so that the product of these two numbers did not overflow a
1519        32-bit integer. However, it turns out that people want to use groups that
1520        are longer than 30,000 bytes (though not repeat them that many times).
1521        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1522        made it possible to implement the integer overflow checks in a much more
1523        dynamic way, which I have now done. The artificial limitation on group
1524        length has been removed - we now have only the limit on the total length of
1525        the compiled pattern, which depends on the LINK_SIZE setting.
1526    
1527    10. Fixed a bug in the documentation for get/copy named substring when
1528        duplicate names are permitted. If none of the named substrings are set, the
1529        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1530        empty string.
1531    
1532    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1533        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1534        because the ] is interpreted as the first data character and the
1535        terminating ] is not found. PCRE has been made compatible with Perl in this
1536        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1537        cause memory overwriting.
1538    
1539    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1540        string has been matched (to stop an infinite loop). It was not recognizing
1541        a conditional subpattern that could match an empty string if that
1542        subpattern was within another subpattern. For example, it looped when
1543        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1544        condition was not nested. This bug has been fixed.
1545    
1546    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1547        past the start of the subject in the presence of bytes with the top bit
1548        set, for example "\x8aBCD".
1549    
1550    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1551        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1552    
1553    14. Optimized (?!) to (*FAIL).
1554    
1555    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1556        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1557        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1558        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1559        does: it's just the validity check that is more restrictive.
1560    
1561    16. Inserted checks for integer overflows during escape sequence (backslash)
1562        processing, and also fixed erroneous offset values for syntax errors during
1563        backslash processing.
1564    
1565    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1566        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1567    
1568    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1569        caused an overrun.
1570    
1571    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1572        something other than just ASCII characters) inside a group that had an
1573        unlimited repeat caused a loop at compile time (while checking to see
1574        whether the group could match an empty string).
1575    
1576    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1577        [\P{Any}] did so. (Error in the code for printing property names.)
1578    
1579    21. An orphan \E inside a character class could cause a crash.
1580    
1581    22. A repeated capturing bracket such as (A)? could cause a wild memory
1582        reference during compilation.
1583    
1584    23. There are several functions in pcre_compile() that scan along a compiled
1585        expression for various reasons (e.g. to see if it's fixed length for look
1586        behind). There were bugs in these functions when a repeated \p or \P was
1587        present in the pattern. These operators have additional parameters compared
1588        with \d, etc, and these were not being taken into account when moving along
1589        the compiled data. Specifically:
1590    
1591        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1592            length.
1593    
1594        (b) An item such as \pL+ within a repeated group could cause crashes or
1595            loops.
1596    
1597        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1598            "reference to non-existent subpattern" error.
1599    
1600        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1601    
1602    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1603        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1604    
1605    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1606    
1607    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1608        character were causing crashes (broken optimization).
1609    
1610    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1611        \p or \P) caused a compile-time loop.
1612    
1613    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1614        break. For example, the unstudied pattern [\r\n]A does not match the string
1615        "\r\nA" because change 7.0/46 below moves the current point on by two
1616        characters after failing to match at the start. However, the pattern \nA
1617        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1618        the same is true. There doesn't seem any very clean way out of this, but
1619        what I have chosen to do makes the common cases work: PCRE now takes note
1620        of whether there can be an explicit match for \r or \n anywhere in the
1621        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1622        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1623        pattern has explicit CR or LF references.
1624    
1625    29. Added (*CR) etc for changing newline setting at start of pattern.
1626    
1627    
1628    Version 7.2 19-Jun-07
1629    ---------------------
1630    
1631     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1632        which is apparently normally available under Windows.
1633    
1634     2. Re-jig the pcregrep tests with different newline settings in an attempt
1635        to make them independent of the local environment's newline setting.
1636    
1637     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1638    
1639     4. Some of the "internals" tests were previously cut out when the link size
1640        was not 2, because the output contained actual offsets. The recent new
1641        "Z" feature of pcretest means that these can be cut out, making the tests
1642        usable with all link sizes.
1643    
1644     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1645        stack recursion. This gives a massive performance boost under BSD, but just
1646        a small improvement under Linux. However, it saves one field in the frame
1647        in all cases.
1648    
1649     6. Added more features from the forthcoming Perl 5.10:
1650    
1651        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1652            recursion call. It refers to the nth most recently opened parentheses.
1653    
1654        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1655            to be opened parentheses.
1656    
1657        (c) Conditions that refer to capturing parentheses can be specified
1658            relatively, for example, (?(-2)... or (?(+3)...
1659    
1660        (d) \K resets the start of the current match so that everything before
1661            is not part of it.
1662    
1663        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1664    
1665        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1666            reference syntax.
1667    
1668        (g) (?| introduces a group in which the numbering of parentheses in each
1669            alternative starts with the same number.
1670    
1671        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1672    
1673     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1674        PCRE_INFO_JCHANGED.
1675    
1676     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1677        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1678        for detecting groups that can match an empty string.
1679    
1680     9. A pattern with a very large number of alternatives (more than several
1681        hundred) was running out of internal workspace during the pre-compile
1682        phase, where pcre_compile() figures out how much memory will be needed. A
1683        bit of new cunning has reduced the workspace needed for groups with
1684        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1685        workspace instead of running out of the 4096 that are available.
1686    
1687    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1688    
1689    11. Applied patch from Google to remove an optimization that didn't quite work.
1690        The report of the bug said:
1691    
1692          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1693          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1694          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1695    
1696    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1697        it matched the wrong number of bytes.
1698    
1699    
1700    Version 7.1 24-Apr-07
1701    ---------------------
1702    
1703     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1704        that is more "standard", making use of automake and other Autotools. There
1705        is some re-arrangement of the files and adjustment of comments consequent
1706        on this.
1707    
1708     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1709        for recursive directory scanning broke on some systems because the files
1710        are not scanned in any specific order and on different systems the order
1711        was different. A call to "sort" has been inserted into RunGrepTest for the
1712        approprate test as a short-term fix. In the longer term there may be an
1713        alternative.
1714    
1715     3. I had an email from Eric Raymond about problems translating some of PCRE's
1716        man pages to HTML (despite the fact that I distribute HTML pages, some
1717        people do their own conversions for various reasons). The problems
1718        concerned the use of low-level troff macros .br and .in. I have therefore
1719        removed all such uses from the man pages (some were redundant, some could
1720        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1721        HTML has been updated to handle .nf/.fi and to complain if it encounters
1722        .br or .in.
1723    
1724     4. Updated comments in configure.ac that get placed in config.h.in and also
1725        arranged for config.h to be included in the distribution, with the name
1726        config.h.generic, for the benefit of those who have to compile without
1727        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1728    
1729     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1730        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1731        makevp.bat for the current PCRE, using the additional files
1732        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1733    
1734     6. A Windows user reported a minor discrepancy with test 2, which turned out
1735        to be caused by a trailing space on an input line that had got lost in his
1736        copy. The trailing space was an accident, so I've just removed it.
1737    
1738     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1739        that is needed.
1740    
1741     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1742        as "const" (a) because they are and (b) because it helps the PHP
1743        maintainers who have recently made a script to detect big data structures
1744        in the php code that should be moved to the .rodata section. I remembered
1745        to update Builducptable as well, so it won't revert if ucptable.h is ever
1746        re-created.
1747    
1748     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1749        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1750        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1751        support is not required. This saves 1.5-2K of code, which is important in
1752        some applications.
1753    
1754        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1755        so as not to refer to the tables, even though these functions will never be
1756        called when UTF-8 support is disabled. Otherwise there are problems with a
1757        shared library.
1758    
1759    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1760    
1761        (a) It was defining its arguments as char * instead of void *.
1762    
1763        (b) It was assuming that all moves were upwards in memory; this was true
1764            a long time ago when I wrote it, but is no longer the case.
1765    
1766        The emulated memove() is provided for those environments that have neither
1767        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1768        is clearly not the case, as these two bugs were recently reported.
1769    
1770    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1771        and Detrail to create the HTML documentation, the .txt form of the man
1772        pages, and it removes trailing spaces from listed files. It also creates
1773        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1774        case, it wraps all the #defines with #ifndefs. This script should be run
1775        before "make dist".
1776    
1777    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1778        with Unicode property support.
1779    
1780        (a) For a maximizing quantifier, if the two different cases of the
1781            character were of different lengths in their UTF-8 codings (there are
1782            some cases like this - I found 11), and the matching function had to
1783            back up over a mixture of the two cases, it incorrectly assumed they
1784            were both the same length.
1785    
1786        (b) When PCRE was configured to use the heap rather than the stack for
1787            recursion during matching, it was not correctly preserving the data for
1788            the other case of a UTF-8 character when checking ahead for a match
1789            while processing a minimizing repeat. If the check also involved
1790            matching a wide character, but failed, corruption could cause an
1791            erroneous result when trying to check for a repeat of the original
1792            character.
1793    
1794    13. Some tidying changes to the testing mechanism:
1795    
1796        (a) The RunTest script now detects the internal link size and whether there
1797            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1798            values substituted by "configure". (The RunGrepTest script already did
1799            this for UTF-8.) The configure.ac script no longer substitutes the
1800            relevant variables.
1801    
1802        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1803            with length and offset values. This means that the output is different
1804            for different internal link sizes. Test 2 is skipped for link sizes
1805            other than 2 because of this, bypassing the problem. Unfortunately,
1806            there was also a test in test 3 (the locale tests) that used /B and
1807            failed for link sizes other than 2. Rather than cut the whole test out,
1808            I have added a new /Z option to pcretest that replaces the length and
1809            offset values with spaces. This is now used to make test 3 independent
1810            of link size. (Test 2 will be tidied up later.)
1811    
1812    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1813        segmentation fault instead of returning the appropriate error message.
1814    
1815    15. In multiline mode when the newline sequence was set to "any", the pattern
1816        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1817        This doesn't seem right; it now treats the CRLF combination as the line
1818        ending, and so does not match in that case. It's only a pattern such as ^$
1819        that would hit this one: something like ^ABC$ would have failed after \r
1820        and then tried again after \r\n.
1821    
1822    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1823        in an attempt to make files that differ only in their line terminators
1824        compare equal. This works on Linux.
1825    
1826    17. Under certain error circumstances pcregrep might try to free random memory
1827        as it exited. This is now fixed, thanks to valgrind.
1828    
1829    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1830        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1831        was because its rules for how to advance for /g after matching an empty
1832        string at the end of a line did not allow for this case. They now check for
1833        it specially.
1834    
1835    20. pcretest is supposed to handle patterns and data of any length, by
1836        extending its buffers when necessary. It was getting this wrong when the
1837        buffer for a data line had to be extended.
1838    
1839    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1840        CRLF as a newline sequence.
1841    
1842    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1843        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1844        I have nevertheless tidied it up.
1845    
1846    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1847    
1848    24. Added a man page for pcre-config.
1849    
1850    
1851    Version 7.0 19-Dec-06
1852    ---------------------
1853    
1854     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1855        moving to gcc 4.1.1.
1856    
1857     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1858        sys/time.h, which is documented as needed for this function. It doesn't
1859        seem to matter on Linux, but it showed up on some releases of OS X.
1860    
1861     3. It seems that there are systems where bytes whose values are greater than
1862        127 match isprint() in the "C" locale. The "C" locale should be the
1863        default when a C program starts up. In most systems, only ASCII printing
1864        characters match isprint(). This difference caused the output from pcretest
1865        to vary, making some of the tests fail. I have changed pcretest so that:
1866    
1867        (a) When it is outputting text in the compiled version of a pattern, bytes
1868            other than 32-126 are always shown as hex escapes.
1869    
1870        (b) When it is outputting text that is a matched part of a subject string,
1871            it does the same, unless a different locale has been set for the match
1872            (using the /L modifier). In this case, it uses isprint() to decide.
1873    
1874     4. Fixed a major bug that caused incorrect computation of the amount of memory
1875        required for a compiled pattern when options that changed within the
1876        pattern affected the logic of the preliminary scan that determines the
1877        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1878        that the computed length was too small. The symptoms of this bug were
1879        either the PCRE error "internal error: code overflow" from pcre_compile(),
1880        or a glibc crash with a message such as "pcretest: free(): invalid next
1881        size (fast)". Examples of patterns that provoked this bug (shown in
1882        pcretest format) are:
1883    
1884          /(?-x: )/x
1885          /(?x)(?-x: \s*#\s*)/
1886          /((?i)[\x{c0}])/8
1887          /(?i:[\x{c0}])/8
1888    
1889        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1890        is now done differently.
1891    
1892     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1893        wrapper classes; (b) implement a new function in the C++ scanner that is
1894        more efficient than the old way of doing things because it avoids levels of
1895        recursion in the regex matching; (c) add a paragraph to the documentation
1896        for the FullMatch() function.
1897    
1898     6. The escape sequence \n was being treated as whatever was defined as
1899        "newline". Not only was this contrary to the documentation, which states
1900        that \n is character 10 (hex 0A), but it also went horribly wrong when
1901        "newline" was defined as CRLF. This has been fixed.
1902    
1903     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1904        was being set to -1 for the "end of line" case (supposedly a value that no
1905        character can have). Though this value is never used (the check for end of
1906        line is "zero bytes in current character"), it caused compiler complaints.
1907        I've changed it to 0xffffffff.
1908    
1909     8. In pcre_version.c, the version string was being built by a sequence of
1910        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1911        string (as it is for production releases) called a macro with an empty
1912        argument. The C standard says the result of this is undefined. The gcc
1913        compiler treats it as an empty string (which was what was wanted) but it is
1914        reported that Visual C gives an error. The source has been hacked around to
1915        avoid this problem.
1916    
1917     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1918        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1919        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1920        of them did).
1921    
1922    10. Originally, pcretest opened its input and output without "b"; then I was
1923        told that "b" was needed in some environments, so it was added for release
1924        5.0 to both the input and output. (It makes no difference on Unix-like
1925        systems.) Later I was told that it is wrong for the input on Windows. I've
1926        now abstracted the modes into two macros, to make it easier to fiddle with
1927        them, and removed "b" from the input mode under Windows.
1928    
1929    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1930    
1931    12. Added -help and --help to pcretest as an official way of being reminded
1932        of the options.
1933    
1934    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1935        and pcrecpp.cc because they annoy compilers at high warning levels.
1936    
1937    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1938    
1939    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1940        scripts are not C programs :-) and which was not noticed because it works
1941        on Linux.
1942    
1943    16. pcretest is supposed to handle any length of pattern and data line (as one
1944        line or as a continued sequence of lines) by extending its input buffer if
1945        necessary. This feature was broken for very long pattern lines, leading to
1946        a string of junk being passed to pcre_compile() if the pattern was longer
1947        than about 50K.
1948    
1949    17. I have done a major re-factoring of the way pcre_compile() computes the
1950        amount of memory needed for a compiled pattern. Previously, there was code
1951        that made a preliminary scan of the pattern in order to do this. That was
1952        OK when PCRE was new, but as the facilities have expanded, it has become
1953        harder and harder to keep it in step with the real compile phase, and there
1954        have been a number of bugs (see for example, 4 above). I have now found a
1955        cunning way of running the real compile function in a "fake" mode that
1956        enables it to compute how much memory it would need, while actually only
1957        ever using a few hundred bytes of working memory and without too many
1958        tests of the mode. This should make future maintenance and development
1959        easier. A side effect of this work is that the limit of 200 on the nesting
1960        depth of parentheses has been removed (though this was never a serious
1961        limitation, I suspect). However, there is a downside: pcre_compile() now
1962        runs more slowly than before (30% or more, depending on the pattern). I
1963        hope this isn't a big issue. There is no effect on runtime performance.
1964    
1965    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1966        newline (only possible for the last line of a file) and it was a
1967        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1968    
1969    19. Added additional timing features to pcretest. (1) The -tm option now times
1970        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1971        separate command line item, by a number that specifies the number of
1972        repeats to use when timing. The default is 50000; this gives better
1973        precision, but takes uncomfortably long for very large patterns.
1974    
1975    20. Extended pcre_study() to be more clever in cases where a branch of a
1976        subpattern has no definite first character. For example, (a*|b*)[cd] would
1977        previously give no result from pcre_study(). Now it recognizes that the
1978        first character must be a, b, c, or d.
1979    
1980    21. There was an incorrect error "recursive call could loop indefinitely" if
1981        a subpattern (or the entire pattern) that was being tested for matching an
1982        empty string contained only one non-empty item after a nested subpattern.
1983        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1984        incorrectly, because the \d was being skipped in the check.
1985    
1986    22. The pcretest program now has a new pattern option /B and a command line
1987        option -b, which is equivalent to adding /B to every pattern. This causes
1988        it to show the compiled bytecode, without the additional information that
1989        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1990        is the same as /B/I).
1991    
1992    23. A new optimization is now able automatically to treat some sequences such
1993        as a*b as a*+b. More specifically, if something simple (such as a character
1994        or a simple class like \d) has an unlimited quantifier, and is followed by
1995        something that cannot possibly match the quantified thing, the quantifier
1996        is automatically "possessified".
1997    
1998    24. A recursive reference to a subpattern whose number was greater than 39
1999        went wrong under certain circumstances in UTF-8 mode. This bug could also
2000        have affected the operation of pcre_study().
2001    
2002    25. Realized that a little bit of performance could be had by replacing
2003        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2004    
2005    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2006    
2007    27. Possessive quantifiers such as a++ were previously implemented by turning
2008        them into atomic groups such as ($>a+). Now they have their own opcodes,
2009        which improves performance. This includes the automatically created ones
2010        from 23 above.
2011    
2012    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2013        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2014        the first matched character to be a colon. This applied both to named and
2015        numbered groups.
2016    
2017    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2018    
2019    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2020        building PCRE on Windows easier, so I have included it in the distribution.
2021    
2022    31. There is now a check in pcretest against a ridiculously large number being
2023        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2024        loop, the loop is abandoned.
2025    
2026    32. Forward references to subpatterns in conditions such as (?(2)...) where
2027        subpattern 2 is defined later cause pcre_compile() to search forwards in
2028        the pattern for the relevant set of parentheses. This search went wrong
2029        when there were unescaped parentheses in a character class, parentheses
2030        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2031    
2032    33. "Subroutine" calls and backreferences were previously restricted to
2033        referencing subpatterns earlier in the regex. This restriction has now
2034        been removed.
2035    
2036    34. Added a number of extra features that are going to be in Perl 5.10. On the
2037        whole, these are just syntactic alternatives for features that PCRE had
2038        previously implemented using the Python syntax or my own invention. The
2039        other formats are all retained for compatibility.
2040    
2041        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2042            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2043            also .NET compatible.
2044    
2045        (b) A recursion or subroutine call to a named group can now be defined as
2046            (?&name) as well as (?P>name).
2047    
2048        (c) A backreference to a named group can now be defined as \k<name> or
2049            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2050            5.10, are also .NET compatible.
2051    
2052        (d) A conditional reference to a named group can now use the syntax
2053            (?(<name>) or (?('name') as well as (?(name).
2054    
2055        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2056            groups (named and numbered) that are never evaluated inline, but can be
2057            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2058            is always false. There may be only one alternative in such a group.
2059    
2060        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2061            as the simple (?(R). The condition is true only if the most recent
2062            recursion is that of the given number or name. It does not search out
2063            through the entire recursion stack.
2064    
2065        (g) The escape \gN or \g{N} has been added, where N is a positive or
2066            negative number, specifying an absolute or relative reference.
2067    
2068    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2069        some "unreachable code" warnings.
2070    
2071    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2072        things, this adds five new scripts.
2073    
2074    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2075        There were also incompatibilities regarding the handling of \Q..\E inside
2076        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2077        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2078    
2079    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2080        matches an empty string, and forcibly breaks the loop. There were bugs in
2081        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2082        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2083        separate and independent bugs (that affected different cases) have been
2084        fixed.
2085    
2086    39. Refactored the code to abolish the use of different opcodes for small
2087        capturing bracket numbers. This is a tidy that I avoided doing when I
2088        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2089        The new approach is not only tidier, it makes it possible to reduce the
2090        memory needed to fix the previous bug (38).
2091    
2092    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2093        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2094        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2095        mode.
2096    
2097    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2098        report.
2099    
2100    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2101        copy construction and assignment in the C++ wrapper.
2102    
2103    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2104        couple of bugs that could have given wrong results in the "--newline=crlf"
2105        case.
2106    
2107    44. Added a number of casts and did some reorganization of signed/unsigned int
2108        variables following suggestions from Dair Grant. Also renamed the variable
2109        "this" as "item" because it is a C++ keyword.
2110    
2111    45. Arranged for dftables to add
2112    
2113          #include "pcre_internal.h"
2114    
2115        to pcre_chartables.c because without it, gcc 4.x may remove the array
2116        definition from the final binary if PCRE is built into a static library and
2117        dead code stripping is activated.
2118    
2119    46. For an unanchored pattern, if a match attempt fails at the start of a
2120        newline sequence, and the newline setting is CRLF or ANY, and the next two
2121        characters are CRLF, advance by two characters instead of one.
2122    
2123    
2124    Version 6.7 04-Jul-06
2125    ---------------------
2126    
2127     1. In order to handle tests when input lines are enormously long, pcretest has
2128        been re-factored so that it automatically extends its buffers when
2129        necessary. The code is crude, but this _is_ just a test program. The
2130        default size has been increased from 32K to 50K.
2131    
2132     2. The code in pcre_study() was using the value of the re argument before
2133        testing it for NULL. (Of course, in any sensible call of the function, it
2134        won't be NULL.)
2135    
2136     3. The memmove() emulation function in pcre_internal.h, which is used on
2137        systems that lack both memmove() and bcopy() - that is, hardly ever -
2138        was missing a "static" storage class specifier.
2139    
2140     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2141        containing an extended class (one that cannot be represented by a bitmap
2142        because it contains high-valued characters or Unicode property items, e.g.
2143        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2144        pattern, but PCRE should not loop if you do not (it no longer does).
2145        [Detail: two cases were found: (a) a repeated subpattern containing an
2146        extended class; (b) a recursive reference to a subpattern that followed a
2147        previous extended class. It wasn't skipping over the extended class
2148        correctly when UTF-8 mode was not set.]
2149    
2150     5. A negated single-character class was not being recognized as fixed-length
2151        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2152        compile error "lookbehind assertion is not fixed length".
2153    
2154     6. The RunPerlTest auxiliary script was showing an unexpected difference
2155        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2156        write a Perl script that can interpret lines of an input file either as
2157        byte characters or as UTF-8, which is what "perltest" was being required to
2158        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2159        can't do is switch easily at run time between having the "use utf8;" pragma
2160        or not. In the end, I fudged it by using the RunPerlTest script to insert
2161        "use utf8;" explicitly for the UTF-8 tests.
2162    
2163     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2164        the end of the subject string, contrary to the documentation and to what
2165        Perl does. This was true of both matching functions. Now it matches only at
2166        the start of the subject and immediately after *internal* newlines.
2167    
2168     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2169        a pointer to an int instead of a pointer to an unsigned long int. This
2170        caused problems on 64-bit systems.
2171    
2172     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2173        instance of the 'standard' template library not being so standard".
2174    
2175    10. There was no check on the number of named subpatterns nor the maximum
2176        length of a subpattern name. The product of these values is used to compute
2177        the size of the memory block for a compiled pattern. By supplying a very
2178        long subpattern name and a large number of named subpatterns, the size
2179        computation could be caused to overflow. This is now prevented by limiting
2180        the length of names to 32 characters, and the number of named subpatterns
2181        to 10,000.
2182    
2183    11. Subpatterns that are repeated with specific counts have to be replicated in
2184        the compiled pattern. The size of memory for this was computed from the
2185        length of the subpattern and the repeat count. The latter is limited to
2186        65535, but there was no limit on the former, meaning that integer overflow
2187        could in principle occur. The compiled length of a repeated subpattern is
2188        now limited to 30,000 bytes in order to prevent this.
2189    
2190    12. Added the optional facility to have named substrings with the same name.
2191    
2192    13. Added the ability to use a named substring as a condition, using the
2193        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2194        are numbers (not recommended). Forward references are permitted.
2195    
2196    14. Added forward references in named backreferences (if you see what I mean).
2197    
2198    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2199        pattern could run off the end of the subject. For example, the pattern
2200        "(?s)(.{1,5})"8 did this with the subject "ab".
2201    
2202    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2203        PCRE_CASELESS was set when matching characters that were quantified with ?
2204        or *.
2205    
2206    17. A character class other than a single negated character that had a minimum
2207        but no maximum quantifier - for example [ab]{6,} - was not handled
2208        correctly by pce_dfa_exec(). It would match only one character.
2209    
2210    18. A valid (though odd) pattern that looked like a POSIX character
2211        class but used an invalid character after [ (for example [[,abc,]]) caused
2212        pcre_compile() to give the error "Failed: internal error: code overflow" or
2213        in some cases to crash with a glibc free() error. This could even happen if
2214        the pattern terminated after [[ but there just happened to be a sequence of
2215        letters, a binary zero, and a closing ] in the memory that followed.
2216    
2217    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2218        over the years. Originally (before any Unicode support), just the bottom 8
2219        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2220        output from "man perlunicode" includes this:
2221    
2222          The regular expression compiler produces polymorphic opcodes.  That
2223          is, the pattern adapts to the data and automatically switches to
2224          the Unicode character scheme when presented with Unicode data--or
2225          instead uses a traditional byte scheme when presented with byte
2226          data.
2227    
2228        Sadly, a wide octal escape does not cause a switch, and in a string with
2229        no other multibyte characters, these octal escapes are treated as before.
2230        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2231        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2232        Unicode string.
2233    
2234        I have not perpetrated such confusion in PCRE. Up till now, it took just
2235        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2236        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2237        translate to the appropriate multibyte character.
2238    
2239    29. Applied some refactoring to reduce the number of warnings from Microsoft
2240        and Borland compilers. This has included removing the fudge introduced
2241        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2242        a warning about an unused variable.
2243    
2244    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2245        characters since release 4.0, because Perl (from release 5.004) does not.
2246        [Or at least, is documented not to: some releases seem to be in conflict
2247        with the documentation.] However, when a pattern was studied with
2248        pcre_study() and all its branches started with \s, PCRE still included VT
2249        as a possible starting character. Of course, this did no harm; it just
2250        caused an unnecessary match attempt.
2251    
2252    22. Removed a now-redundant internal flag bit that recorded the fact that case
2253        dependency changed within the pattern. This was once needed for "required
2254        byte" processing, but is no longer used. This recovers a now-scarce options
2255        bit. Also moved the least significant internal flag bit to the most-
2256        significant bit of the word, which was not previously used (hangover from
2257        the days when it was an int rather than a uint) to free up another bit for
2258        the future.
2259    
2260    23. Added support for CRLF line endings as well as CR and LF. As well as the
2261        default being selectable at build time, it can now be changed at runtime
2262        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2263        specify that it is scanning data with non-default line endings.
2264    
2265    24. Changed the definition of CXXLINK to make it agree with the definition of
2266        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2267    
2268    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2269        recursions. This makes a big different to stack usage for some patterns.
2270    
2271    26. If a subpattern containing a named recursion or subroutine reference such
2272        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2273        the space required for the compiled pattern went wrong and gave too small a
2274        value. Depending on the environment, this could lead to "Failed: internal
2275        error: code overflow at offset 49" or "glibc detected double free or
2276        corruption" errors.
2277    
2278    27. Applied patches from Google (a) to support the new newline modes and (b) to
2279        advance over multibyte UTF-8 characters in GlobalReplace.
2280    
2281    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2282        difference for some implementation of PCRE in some Windows version.
2283    
2284    29. Added some extra testing facilities to pcretest:
2285    
2286        \q<number>   in a data line sets the "match limit" value
2287        \Q<number>   in a data line sets the "match recursion limt" value
2288        -S <number>  sets the stack size, where <number> is in megabytes
2289    
2290        The -S option isn't available for Windows.
2291    
2292    
2293    Version 6.6 06-Feb-06
2294    ---------------------
2295    
2296     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2297        in pcreposix.h. I have copied the definition from pcre.h.
2298    
2299     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2300        because pcre.h is no longer a built file.
2301    
2302     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2303        not normally included in the compiled code.
2304    
2305    
2306  Version 6.5 01-Feb-06  Version 6.5 01-Feb-06
2307  ---------------------  ---------------------
2308    

Legend:
Removed from v.87  
changed lines
  Added in v.663

  ViewVC Help
Powered by ViewVC 1.1.5