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

Diff of /code/trunk/ChangeLog

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

revision 218 by ph10, Thu Aug 16 10:13:23 2007 UTC revision 747 by ph10, Tue Nov 15 17:35:10 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.3 09-Aug-07  Version 8.21
5    ------------
6    
7    1.  Updating the JIT compiler.
8    
9    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10        are added as well.
11    
12    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14        calling _pcre_jit_exec. Some extra comments are added.
15    
16    4.  Mark settings inside atomic groups that do not contain any capturing
17        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18        was introduced by change 18 for 8.20.
19    
20    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21        ECMA-262 standard.
22    
23    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25        This bug was probably introduced by change 9 of 8.13.
26    
27    7.  While fixing 6 above, I noticed that a number of other items were being
28        incorrectly rejected as "not fixed length". This arose partly because newer
29        opcodes had not been added to the fixed-length checking code. I have (a)
30        corrected the bug and added tests for these items, and (b) arranged for an
31        error to occur if an unknown opcode is encountered while checking for fixed
32        length instead of just assuming "not fixed length". The items that were
33        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36    
37    
38    Version 8.20 21-Oct-2011
39    ------------------------
40    
41    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
42        a POSIX class. After further experiments with Perl, which convinced me that
43        Perl has bugs and confusions, a closing square bracket is no longer allowed
44        in a POSIX name. This bug also affected patterns with classes that started
45        with full stops.
46    
47    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
48        captured substring, but while checking the failing first alternative,
49        substring 1 is temporarily captured. If the output vector supplied to
50        pcre_exec() was not big enough for this capture, the yield of the function
51        was still zero ("insufficient space for captured substrings"). This cannot
52        be totally fixed without adding another stack variable, which seems a lot
53        of expense for a edge case. However, I have improved the situation in cases
54        such as /(a)(b)x|abc/ matched against "abc", where the return code
55        indicates that fewer than the maximum number of slots in the ovector have
56        been set.
57    
58    3.  Related to (2) above: when there are more back references in a pattern than
59        slots in the output vector, pcre_exec() uses temporary memory during
60        matching, and copies in the captures as far as possible afterwards. It was
61        using the entire output vector, but this conflicts with the specification
62        that only 2/3 is used for passing back captured substrings. Now it uses
63        only the first 2/3, for compatibility. This is, of course, another edge
64        case.
65    
66    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
67        main code base, and can be used by building with --enable-jit. When this is
68        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
69        runtime --no-jit option is given.
70    
71    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
72        ovector, the return from the function was zero, implying that there were
73        other matches that did not fit. The correct "exactly full" value is now
74        returned.
75    
76    6.  If a subpattern that was called recursively or as a subroutine contained
77        (*PRUNE) or any other control that caused it to give a non-standard return,
78        invalid errors such as "Error -26 (nested recursion at the same subject
79        position)" or even infinite loops could occur.
80    
81    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
82        computing the minimum length on reaching *ACCEPT, and so ended up with the
83        wrong value of 1 rather than 0. Further investigation indicates that
84        computing a minimum subject length in the presence of *ACCEPT is difficult
85        (think back references, subroutine calls), and so I have changed the code
86        so that no minimum is registered for a pattern that contains *ACCEPT.
87    
88    8.  If (*THEN) was present in the first (true) branch of a conditional group,
89        it was not handled as intended. [But see 16 below.]
90    
91    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
92        Sheri Pierce.
93    
94    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
95        the first byte in a match must be "a".
96    
97    11. Change 17 for 8.13 increased the recursion depth for patterns like
98        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
99        pattern contains any instances of (*THEN). If it does not, the old
100        optimizations are restored. It would be nice to do this on a per-group
101        basis, but at the moment that is not feasible.
102    
103    12. In some environments, the output of pcretest -C is CRLF terminated. This
104        broke RunTest's code that checks for the link size. A single white space
105        character after the value is now allowed for.
106    
107    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
108        For "fr", it uses the Windows-specific input and output files.
109    
110    14. If (*THEN) appeared in a group that was called recursively or as a
111        subroutine, it did not work as intended. [But see next item.]
112    
113    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
114        pattern fragments (but not containing any | characters). If A and B are
115        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
116        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
117        D. In other words, Perl considers parentheses that do not contain any |
118        characters to be part of a surrounding alternative, whereas PCRE was
119        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
120        differently. PCRE now behaves in the same way as Perl, except in the case
121        of subroutine/recursion calls such as (?1) which have in any case always
122        been different (but PCRE had them first :-).
123    
124    16. Related to 15 above: Perl does not treat the | in a conditional group as
125        creating alternatives. Such a group is treated in the same way as an
126        ordinary group without any | characters when processing (*THEN). PCRE has
127        been changed to match Perl's behaviour.
128    
129    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
130        RunGrepTest script failed.
131    
132    18. Change 22 for version 13 caused atomic groups to use more stack. This is
133        inevitable for groups that contain captures, but it can lead to a lot of
134        stack use in large patterns. The old behaviour has been restored for atomic
135        groups that do not contain any capturing parentheses.
136    
137    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
138        suppress the check for a minimum subject length at run time. (If it was
139        given to pcre_exec() or pcre_dfa_exec() it did work.)
140    
141    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
142        fail to work when decoding hex characters in data strings in EBCDIC
143        environments.
144    
145    21. It appears that in at least one Mac OS environment, the isxdigit() function
146        is implemented as a macro that evaluates to its argument more than once,
147        contravening the C 90 Standard (I haven't checked a later standard). There
148        was an instance in pcretest which caused it to go wrong when processing
149        \x{...} escapes in subject strings. The has been rewritten to avoid using
150        things like p++ in the argument of isxdigit().
151    
152    
153    Version 8.13 16-Aug-2011
154    ------------------------
155    
156    1.  The Unicode data tables have been updated to Unicode 6.0.0.
157    
158    2.  Two minor typos in pcre_internal.h have been fixed.
159    
160    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
161        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
162        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
163    
164    4.  There were a number of related bugs in the code for matching backrefences
165        caselessly in UTF-8 mode when codes for the characters concerned were
166        different numbers of bytes. For example, U+023A and U+2C65 are an upper
167        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
168        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
169        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
170        2-byte code at the end of the subject (it thought there wasn't enough data
171        left).
172    
173    5.  Comprehensive information about what went wrong is now returned by
174        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
175        as the output vector has at least 2 elements. The offset of the start of
176        the failing character and a reason code are placed in the vector.
177    
178    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
179        now returned is for the first byte of the failing character, instead of the
180        last byte inspected. This is an incompatible change, but I hope it is small
181        enough not to be a problem. It makes the returned offset consistent with
182        pcre_exec() and pcre_dfa_exec().
183    
184    7.  pcretest now gives a text phrase as well as the error number when
185        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
186        failure, the offset and reason code are output.
187    
188    8.  When \R was used with a maximizing quantifier it failed to skip backwards
189        over a \r\n pair if the subsequent match failed. Instead, it just skipped
190        back over a single character (\n). This seems wrong (because it treated the
191        two characters as a single entity when going forwards), conflicts with the
192        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
193        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
194        has been changed.
195    
196    9.  Some internal refactoring has changed the processing so that the handling
197        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
198        time (the PCRE_DOTALL option was changed this way some time ago: version
199        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
200        which was always a bit of a fudge. It also means that there is one less
201        argument for the match() function, which reduces its stack requirements
202        slightly. This change also fixes an incompatibility with Perl: the pattern
203        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
204    
205    10. More internal refactoring has drastically reduced the number of recursive
206        calls to match() for possessively repeated groups such as (abc)++ when
207        using pcre_exec().
208    
209    11. While implementing 10, a number of bugs in the handling of groups were
210        discovered and fixed:
211    
212        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
213        (a|)*(?1) gave a compile-time internal error.
214        ((a|)+)+  did not notice that the outer group could match an empty string.
215        (^a|^)+   was not marked as anchored.
216        (.*a|.*)+ was not marked as matching at start or after a newline.
217    
218    12. Yet more internal refactoring has removed another argument from the match()
219        function. Special calls to this function are now indicated by setting a
220        value in a variable in the "match data" data block.
221    
222    13. Be more explicit in pcre_study() instead of relying on "default" for
223        opcodes that mean there is no starting character; this means that when new
224        ones are added and accidentally left out of pcre_study(), testing should
225        pick them up.
226    
227    14. The -s option of pcretest has been documented for ages as being an old
228        synonym of -m (show memory usage). I have changed it to mean "force study
229        for every regex", that is, assume /S for every regex. This is similar to -i
230        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
231        using it. It makes it easier to run collections of tests with and without
232        study enabled, and thereby test pcre_study() more easily. All the standard
233        tests are now run with and without -s (but some patterns can be marked as
234        "never study" - see 20 below).
235    
236    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
237        restoration of the capturing data to the outer values was not happening
238        correctly.
239    
240    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
241        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
242        pattern had matched an empty string, and so incorrectly returned a no
243        match.
244    
245    17. There was optimizing code for the last branch of non-capturing parentheses,
246        and also for the obeyed branch of a conditional subexpression, which used
247        tail recursion to cut down on stack usage. Unfortunately, now that there is
248        the possibility of (*THEN) occurring in these branches, tail recursion is
249        no longer possible because the return has to be checked for (*THEN). These
250        two optimizations have therefore been removed. [But see 8.20/11 above.]
251    
252    18. If a pattern containing \R was studied, it was assumed that \R always
253        matched two bytes, thus causing the minimum subject length to be
254        incorrectly computed because \R can also match just one byte.
255    
256    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
257        was incorrectly computed.
258    
259    20. If /S is present twice on a test pattern in pcretest input, it now
260        *disables* studying, thereby overriding the use of -s on the command line
261        (see 14 above). This is necessary for one or two tests to keep the output
262        identical in both cases.
263    
264    21. When (*ACCEPT) was used in an assertion that matched an empty string and
265        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
266    
267    22. When an atomic group that contained a capturing parenthesis was
268        successfully matched, but the branch in which it appeared failed, the
269        capturing was not being forgotten if a higher numbered group was later
270        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
271        group 1 to "a", when in fact it should be unset. This applied to multi-
272        branched capturing and non-capturing groups, repeated or not, and also to
273        positive assertions (capturing in negative assertions does not happen
274        in PCRE) and also to nested atomic groups.
275    
276    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
277        subject after a captured substring, to make it easier to tell which of a
278        number of identical substrings has been captured.
279    
280    24. The way atomic groups are processed by pcre_exec() has been changed so that
281        if they are repeated, backtracking one repetition now resets captured
282        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
283        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
284        "aaa". Previously, it would have been "a". As part of this code
285        refactoring, the way recursive calls are handled has also been changed.
286    
287    25. If an assertion condition captured any substrings, they were not passed
288        back unless some other capturing happened later. For example, if
289        (?(?=(a))a) was matched against "a", no capturing was returned.
290    
291    26. When studying a pattern that contained subroutine calls or assertions,
292        the code for finding the minimum length of a possible match was handling
293        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
294        group 1 called group 2 while simultaneously a separate group 2 called group
295        1). A stack overflow occurred in this case. I have fixed this by limiting
296        the recursion depth to 10.
297    
298    27. Updated RunTest.bat in the distribution to the version supplied by Tom
299        Fortmann. This supports explicit test numbers on the command line, and has
300        argument validation and error reporting.
301    
302    28. An instance of \X with an unlimited repeat could fail if at any point the
303        first character it looked at was a mark character.
304    
305    29. Some minor code refactoring concerning Unicode properties and scripts
306        should reduce the stack requirement of match() slightly.
307    
308    30. Added the '=' option to pcretest to check the setting of unused capturing
309        slots at the end of the pattern, which are documented as being -1, but are
310        not included in the return count.
311    
312    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
313        compiled something random. Now it gives a compile-time error (as does
314        Perl).
315    
316    32. A *MARK encountered during the processing of a positive assertion is now
317        recorded and passed back (compatible with Perl).
318    
319    33. If --only-matching or --colour was set on a pcregrep call whose pattern
320        had alternative anchored branches, the search for a second match in a line
321        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
322        matched the line "0102" twice. The same bug affected patterns that started
323        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
324        twice.
325    
326    34. Previously, PCRE did not allow quantification of assertions. However, Perl
327        does, and because of capturing effects, quantifying parenthesized
328        assertions may at times be useful. Quantifiers are now allowed for
329        parenthesized assertions.
330    
331    35. A minor code tidy in pcre_compile() when checking options for \R usage.
332    
333    36. \g was being checked for fancy things in a character class, when it should
334        just be a literal "g".
335    
336    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
337        appearance of a nested POSIX class supersedes an apparent external class.
338        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
339        unescaped square brackets may also appear as part of class names. For
340        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
341        more like Perl. (But see 8.20/1 above.)
342    
343    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
344        was because it thought it was \N{name}, which is not supported).
345    
346    39. Add minix to OS list not supporting the -S option in pcretest.
347    
348    40. PCRE tries to detect cases of infinite recursion at compile time, but it
349        cannot analyze patterns in sufficient detail to catch mutual recursions
350        such as ((?1))((?2)). There is now a runtime test that gives an error if a
351        subgroup is called recursively as a subpattern for a second time at the
352        same position in the subject string. In previous releases this might have
353        been caught by the recursion limit, or it might have run out of stack.
354    
355    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
356        happen only once. PCRE was, however incorrectly giving a compile time error
357        "recursive call could loop indefinitely" because it cannot analyze the
358        pattern in sufficient detail. The compile time test no longer happens when
359        PCRE is compiling a conditional subpattern, but actual runaway loops are
360        now caught at runtime (see 40 above).
361    
362    42. It seems that Perl allows any characters other than a closing parenthesis
363        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
364        has been changed to be the same.
365    
366    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
367        as not to get warnings when autogen.sh is called. Also changed
368        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
369    
370    44. To help people who use pcregrep to scan files containing exceedingly long
371        lines, the following changes have been made:
372    
373        (a) The default value of the buffer size parameter has been increased from
374            8K to 20K. (The actual buffer used is three times this size.)
375    
376        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
377            PCRE is built.
378    
379        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
380            to be set at run time.
381    
382        (d) Numerical values in pcregrep options can be followed by K or M, for
383            example --buffer-size=50K.
384    
385        (e) If a line being scanned overflows pcregrep's buffer, an error is now
386            given and the return code is set to 2.
387    
388    45. Add a pointer to the latest mark to the callout data block.
389    
390    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
391        partial match of an empty string instead of no match. This was specific to
392        the use of ".".
393    
394    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
395        complete match instead of a partial match. This bug was dependent on both
396        the PCRE_UTF8 and PCRE_DOTALL options being set.
397    
398    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
399        starting byte set, because \b was not being ignored.
400    
401    
402    Version 8.12 15-Jan-2011
403    ------------------------
404    
405    1.  Fixed some typos in the markup of the man pages, and wrote a script that
406        checks for such things as part of the documentation building process.
407    
408    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
409        --match-limit and --recursion-limit options (added for 8.11). In
410        particular, this made one of the standard tests fail. (The integer value
411        went into the wrong half of a long int.)
412    
413    3.  If the --colour option was given to pcregrep with -v (invert match), it
414        did strange things, either producing crazy output, or crashing. It should,
415        of course, ignore a request for colour when reporting lines that do not
416        match.
417    
418    4.  Another pcregrep bug caused similar problems if --colour was specified with
419        -M (multiline) and the pattern match finished with a line ending.
420    
421    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
422        matched in multiline mode, the following line was shown as part of the
423        match. This seems wrong, so I have changed it.
424    
425    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
426        the check for further matches in the same line (so they could be coloured)
427        to overrun the end of the current line. If another match was found, it was
428        incorrectly shown (and then shown again when found in the next line).
429    
430    7.  If pcregrep was compiled under Windows, there was a reference to the
431        function pcregrep_exit() before it was defined. I am assuming this was
432        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
433        reported by a user. I've moved the definition above the reference.
434    
435    
436    Version 8.11 10-Dec-2010
437    ------------------------
438    
439    1.  (*THEN) was not working properly if there were untried alternatives prior
440        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
441        backtracked to try for "b" instead of moving to the next alternative branch
442        at the same level (in this case, to look for "c"). The Perl documentation
443        is clear that when (*THEN) is backtracked onto, it goes to the "next
444        alternative in the innermost enclosing group".
445    
446    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
447        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
448        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
449        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
450        (*THEN).
451    
452    3.  If \s appeared in a character class, it removed the VT character from
453        the class, even if it had been included by some previous item, for example
454        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
455        of \s, but is part of the POSIX "space" class.)
456    
457    4.  A partial match never returns an empty string (because you can always
458        match an empty string at the end of the subject); however the checking for
459        an empty string was starting at the "start of match" point. This has been
460        changed to the "earliest inspected character" point, because the returned
461        data for a partial match starts at this character. This means that, for
462        example, /(?<=abc)def/ gives a partial match for the subject "abc"
463        (previously it gave "no match").
464    
465    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
466        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
467        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
468        has an implication that the given string is incomplete (because a partial
469        match is preferred over a full match). For this reason, these items now
470        give a partial match in this situation. [Aside: previously, the one case
471        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
472        match rather than a full match, which was wrong by the old rules, but is
473        now correct.]
474    
475    6.  There was a bug in the handling of #-introduced comments, recognized when
476        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
477        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
478        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
479        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
480        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
481        places in pcre_compile().
482    
483    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
484        comments when looking ahead for named forward references to subpatterns,
485        the only newline sequence it recognized was NL. It now handles newlines
486        according to the set newline convention.
487    
488    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
489        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
490        cater for a lack of strerror(). These oversights have been fixed.
491    
492    9.  Added --match-limit and --recursion-limit to pcregrep.
493    
494    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
495    
496    11. When the -o option was used, pcregrep was setting a return code of 1, even
497        when matches were found, and --line-buffered was not being honoured.
498    
499    12. Added an optional parentheses number to the -o and --only-matching options
500        of pcregrep.
501    
502    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
503        can match an empty string. The code to do it in pcretest and pcredemo
504        needed fixing:
505    
506        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
507            only one byte after an empty string match just before CRLF (this case
508            just got forgotten; "any" and "anycrlf" were OK).
509    
510        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
511            mode when an empty string match preceded an ASCII character followed by
512            a non-ASCII character. (The code for advancing by one character rather
513            than one byte was nonsense.)
514    
515        (c) The pcredemo.c sample program did not have any code at all to handle
516            the cases when CRLF is a valid newline sequence.
517    
518    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
519        as a starting offset was within the subject string. There is now a new
520        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
521        negative or greater than the length of the string. In order to test this,
522        pcretest is extended to allow the setting of negative starting offsets.
523    
524    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
525        starting offset points to the beginning of a UTF-8 character was
526        unnecessarily clumsy. I tidied it up.
527    
528    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
529        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
530    
531    17. Nobody had reported that the --include_dir option, which was added in
532        release 7.7 should have been called --include-dir (hyphen, not underscore)
533        for compatibility with GNU grep. I have changed it to --include-dir, but
534        left --include_dir as an undocumented synonym, and the same for
535        --exclude-dir, though that is not available in GNU grep, at least as of
536        release 2.5.4.
537    
538    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
539        characters from a string of bytes) have been redefined so as not to use
540        loops, in order to improve performance in some environments. At the same
541        time, I abstracted some of the common code into auxiliary macros to save
542        repetition (this should not affect the compiled code).
543    
544    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
545        compile-time error is now given if \c is not followed by an ASCII
546        character, that is, a byte less than 128. (In EBCDIC mode, the code is
547        different, and any byte value is allowed.)
548    
549    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
550        START_OPTIMIZE option, which is now allowed at compile time - but just
551        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
552        to pcregrep and other applications that have no direct access to PCRE
553        options. The new /Y option in pcretest sets this option when calling
554        pcre_compile().
555    
556    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
557        back references. Groups containing recursive back references were forced to
558        be atomic by that change, but in the case of named groups, the amount of
559        memory required was incorrectly computed, leading to "Failed: internal
560        error: code overflow". This has been fixed.
561    
562    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
563        pcretest.c, to avoid build problems in some Borland environments.
564    
565    
566    Version 8.10 25-Jun-2010
567    ------------------------
568    
569    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
570        THEN.
571    
572    2.  (*ACCEPT) was not working when inside an atomic group.
573    
574    3.  Inside a character class, \B is treated as a literal by default, but
575        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
576        causes the error). The code is unchanged, but I tidied the documentation.
577    
578    4.  Inside a character class, PCRE always treated \R and \X as literals,
579        whereas Perl faults them if its -w option is set. I have changed PCRE so
580        that it faults them when PCRE_EXTRA is set.
581    
582    5.  Added support for \N, which always matches any character other than
583        newline. (It is the same as "." when PCRE_DOTALL is not set.)
584    
585    6.  When compiling pcregrep with newer versions of gcc which may have
586        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
587        declared with attribute warn_unused_result" were given. Just casting the
588        result to (void) does not stop the warnings; a more elaborate fudge is
589        needed. I've used a macro to implement this.
590    
591    7.  Minor change to pcretest.c to avoid a compiler warning.
592    
593    8.  Added four artifical Unicode properties to help with an option to make
594        \s etc use properties (see next item). The new properties are: Xan
595        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
596    
597    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
598        use Unicode properties. (*UCP) at the start of a pattern can be used to set
599        this option. Modified pcretest to add /W to test this facility. Added
600        REG_UCP to make it available via the POSIX interface.
601    
602    10. Added --line-buffered to pcregrep.
603    
604    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
605        studied, and the match started with a letter with a code point greater than
606        127 whose first byte was different to the first byte of the other case of
607        the letter, the other case of this starting letter was not recognized
608        (#976).
609    
610    12. If a pattern that was studied started with a repeated Unicode property
611        test, for example, \p{Nd}+, there was the theoretical possibility of
612        setting up an incorrect bitmap of starting bytes, but fortunately it could
613        not have actually happened in practice until change 8 above was made (it
614        added property types that matched character-matching opcodes).
615    
616    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
617        possible starting bytes for non-anchored patterns.
618    
619    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
620        \R, and also a number of cases that involve Unicode properties, both
621        explicit and implicit when PCRE_UCP is set.
622    
623    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
624        input, it could crash or give wrong results if characters with values
625        greater than 0xc0 were present in the subject string. (Detail: it assumed
626        UTF-8 input when processing these items.)
627    
628    16. Added a lot of (int) casts to avoid compiler warnings in systems where
629        size_t is 64-bit (#991).
630    
631    17. Added a check for running out of memory when PCRE is compiled with
632        --disable-stack-for-recursion (#990).
633    
634    18. If the last data line in a file for pcretest does not have a newline on
635        the end, a newline was missing in the output.
636    
637    19. The default pcre_chartables.c file recognizes only ASCII characters (values
638        less than 128) in its various bitmaps. However, there is a facility for
639        generating tables according to the current locale when PCRE is compiled. It
640        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
641        characters, are recognized by isspace() and therefore were getting set in
642        these tables, and indeed these tables seem to approximate to ISO 8859. This
643        caused a problem in UTF-8 mode when pcre_study() was used to create a list
644        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
645        which of course cannot start UTF-8 characters. I have changed the code so
646        that only real ASCII characters (less than 128) and the correct starting
647        bytes for UTF-8 encodings are set for characters greater than 127 when in
648        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
649        altogether.)
650    
651    20. Added the /T option to pcretest so as to be able to run tests with non-
652        standard character tables, thus making it possible to include the tests
653        used for 19 above in the standard set of tests.
654    
655    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
656        reference to a subpattern the other side of a comment that contains an
657        opening parenthesis caused either an internal compiling error, or a
658        reference to the wrong subpattern.
659    
660    
661    Version 8.02 19-Mar-2010
662    ------------------------
663    
664    1.  The Unicode data tables have been updated to Unicode 5.2.0.
665    
666    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
667        configured.
668    
669    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
670        original author of that file, following a query about its status.
671    
672    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
673        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
674    
675    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
676        quantifier applied to a forward-referencing subroutine call, could compile
677        incorrect code or give the error "internal error: previously-checked
678        referenced subpattern not found".
679    
680    6.  Both MS Visual Studio and Symbian OS have problems with initializing
681        variables to point to external functions. For these systems, therefore,
682        pcre_malloc etc. are now initialized to local functions that call the
683        relevant global functions.
684    
685    7.  There were two entries missing in the vectors called coptable and poptable
686        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
687        I've fixed the data, and added a kludgy way of testing at compile time that
688        the lengths are correct (equal to the number of opcodes).
689    
690    8.  Following on from 7, I added a similar kludge to check the length of the
691        eint vector in pcreposix.c.
692    
693    9.  Error texts for pcre_compile() are held as one long string to avoid too
694        much relocation at load time. To find a text, the string is searched,
695        counting zeros. There was no check for running off the end of the string,
696        which could happen if a new error number was added without updating the
697        string.
698    
699    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
700    
701    11. \K was not working if it appeared in an atomic group or in a group that
702        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
703        \K is "not well defined" if used in an assertion. PCRE now accepts it if
704        the assertion is positive, but not if it is negative.
705    
706    12. Change 11 fortuitously reduced the size of the stack frame used in the
707        "match()" function of pcre_exec.c by one pointer. Forthcoming
708        implementation of support for (*MARK) will need an extra pointer on the
709        stack; I have reserved it now, so that the stack frame size does not
710        decrease.
711    
712    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
713        item in branch that calls a recursion is a subroutine call - as in the
714        second branch in the above example - was incorrectly given the compile-
715        time error "recursive call could loop indefinitely" because pcre_compile()
716        was not correctly checking the subroutine for matching a non-empty string.
717    
718    14. The checks for overrunning compiling workspace could trigger after an
719        overrun had occurred. This is a "should never occur" error, but it can be
720        triggered by pathological patterns such as hundreds of nested parentheses.
721        The checks now trigger 100 bytes before the end of the workspace.
722    
723    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
724    
725    
726    Version 8.01 19-Jan-2010
727    ------------------------
728    
729    1.  If a pattern contained a conditional subpattern with only one branch (in
730        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
731        computed the wrong minimum data length (which is of course zero for such
732        subpatterns). This could cause incorrect "no match" results.
733    
734    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
735        the pattern is reset in the first branch, pcre_compile() failed with
736        "internal error: code overflow at offset...". This happened only when
737        the reset was to the original external option setting. (An optimization
738        abstracts leading options settings into an external setting, which was the
739        cause of this.)
740    
741    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
742        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
743        assertion pattern did not match (meaning that the assertion was true), it
744        was incorrectly treated as false if the SKIP had been reached during the
745        matching. This also applied to assertions used as conditions.
746    
747    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
748        assertion subpattern, including such a pattern used as a condition,
749        unpredictable results occurred, instead of the error return
750        PCRE_ERROR_DFA_UITEM.
751    
752    5.  The C++ GlobalReplace function was not working like Perl for the special
753        situation when an empty string is matched. It now does the fancy magic
754        stuff that is necessary.
755    
756    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
757        removed. (These were left over from very, very early versions of PCRE.)
758    
759    7.  Some cosmetic changes to the code to make life easier when compiling it
760        as part of something else:
761    
762        (a) Change DEBUG to PCRE_DEBUG.
763    
764        (b) In pcre_compile(), rename the member of the "branch_chain" structure
765            called "current" as "current_branch", to prevent a collision with the
766            Linux macro when compiled as a kernel module.
767    
768        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
769            prevent a collision with the Linux macro when compiled as a kernel
770            module.
771    
772    8.  In pcre_compile() there are some checks for integer overflows that used to
773        cast potentially large values to (double). This has been changed to that
774        when building, a check for int64_t is made, and if it is found, it is used
775        instead, thus avoiding the use of floating point arithmetic. (There is no
776        other use of FP in PCRE.) If int64_t is not found, the fallback is to
777        double.
778    
779    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
780        2005 (difference between two addresses compared to an unsigned value).
781    
782    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
783        custom one, because of the following reported problem in Windows:
784    
785          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
786              under Win32.
787          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
788              therefore missing the function definition.
789          - The compiler thus generates a "C" signature for the test function.
790          - The linker fails to find the "C" function.
791          - PCRE fails to configure if asked to do so against libbz2.
792    
793    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
794        messages were output:
795    
796          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
797          rerunning libtoolize, to keep the correct libtool macros in-tree.
798          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
799    
800        I have done both of these things.
801    
802    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
803        most of the time, it *can* run out if it is given a pattern that contains a
804        runaway infinite recursion. I updated the discussion in the pcrestack man
805        page.
806    
807    13. Now that we have gone to the x.xx style of version numbers, the minor
808        version may start with zero. Using 08 or 09 is a bad idea because users
809        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
810        interpreted as invalid octal numbers. I've updated the previous comment in
811        configure.ac, and also added a check that gives an error if 08 or 09 are
812        used.
813    
814    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
815        causing partial matching to fail when the end of the subject matched \W
816        in a UTF-8 pattern where \W was quantified with a minimum of 3.
817    
818    15. There were some discrepancies between the declarations in pcre_internal.h
819        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
820        their definitions. The declarations used "const uschar *" and the
821        definitions used USPTR. Even though USPTR is normally defined as "const
822        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
823        reported that: "This difference in casting confuses some C++ compilers, for
824        example, SunCC recognizes above declarations as different functions and
825        generates broken code for hbpcre." I have changed the declarations to use
826        USPTR.
827    
828    16. GNU libtool is named differently on some systems. The autogen.sh script now
829        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
830        (FreeBSD).
831    
832    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
833        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
834        comment: "Figure out how to create a longlong from a string: strtoll and
835        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
836        instance, but it only takes 2 args instead of 3!"
837    
838    18. A subtle bug concerned with back references has been fixed by a change of
839        specification, with a corresponding code fix. A pattern such as
840        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
841        refers, was giving matches when it shouldn't. For example, xa=xaaa would
842        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
843        same bug. Such groups have to be quantified to be useful, or contained
844        inside another quantified group. (If there's no repetition, the reference
845        can never match.) The problem arises because, having left the group and
846        moved on to the rest of the pattern, a later failure that backtracks into
847        the group uses the captured value from the final iteration of the group
848        rather than the correct earlier one. I have fixed this in PCRE by forcing
849        any group that contains a reference to itself to be an atomic group; that
850        is, there cannot be any backtracking into it once it has completed. This is
851        similar to recursive and subroutine calls.
852    
853    
854    Version 8.00 19-Oct-09
855    ----------------------
856    
857    1.  The table for translating pcre_compile() error codes into POSIX error codes
858        was out-of-date, and there was no check on the pcre_compile() error code
859        being within the table. This could lead to an OK return being given in
860        error.
861    
862    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
863        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
864        in a Windows environment.
865    
866    3.  The pcregrep --count option prints the count for each file even when it is
867        zero, as does GNU grep. However, pcregrep was also printing all files when
868        --files-with-matches was added. Now, when both options are given, it prints
869        counts only for those files that have at least one match. (GNU grep just
870        prints the file name in this circumstance, but including the count seems
871        more useful - otherwise, why use --count?) Also ensured that the
872        combination -clh just lists non-zero counts, with no names.
873    
874    4.  The long form of the pcregrep -F option was incorrectly implemented as
875        --fixed_strings instead of --fixed-strings. This is an incompatible change,
876        but it seems right to fix it, and I didn't think it was worth preserving
877        the old behaviour.
878    
879    5.  The command line items --regex=pattern and --regexp=pattern were not
880        recognized by pcregrep, which required --regex pattern or --regexp pattern
881        (with a space rather than an '='). The man page documented the '=' forms,
882        which are compatible with GNU grep; these now work.
883    
884    6.  No libpcreposix.pc file was created for pkg-config; there was just
885        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
886    
887    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
888        when UCP support is not needed, by modifying the Python script that
889        generates it from Unicode data files. This should not matter if the module
890        is correctly used as a library, but I received one complaint about 50K of
891        unwanted data. My guess is that the person linked everything into his
892        program rather than using a library. Anyway, it does no harm.
893    
894    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
895        was a minimum greater than 1 for a wide character in a possessive
896        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
897        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
898        character. Chaos in the form of incorrect output or a compiling loop could
899        result.
900    
901    9.  The restrictions on what a pattern can contain when partial matching is
902        requested for pcre_exec() have been removed. All patterns can now be
903        partially matched by this function. In addition, if there are at least two
904        slots in the offset vector, the offset of the earliest inspected character
905        for the match and the offset of the end of the subject are set in them when
906        PCRE_ERROR_PARTIAL is returned.
907    
908    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
909        synonymous with PCRE_PARTIAL, for backwards compatibility, and
910        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
911        and may be more useful for multi-segment matching.
912    
913    11. Partial matching with pcre_exec() is now more intuitive. A partial match
914        used to be given if ever the end of the subject was reached; now it is
915        given only if matching could not proceed because another character was
916        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
917        string "Z", which now yields "no match" instead of "partial match". In the
918        case of pcre_dfa_exec(), "no match" is given if every matching path for the
919        final character ended with (*FAIL).
920    
921    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
922        if the pattern had a "must contain" character that was already found in the
923        earlier partial match, unless partial matching was again requested. For
924        example, with the pattern /dog.(body)?/, the "must contain" character is
925        "g". If the first part-match was for the string "dog", restarting with
926        "sbody" failed. This bug has been fixed.
927    
928    13. The string returned by pcre_dfa_exec() after a partial match has been
929        changed so that it starts at the first inspected character rather than the
930        first character of the match. This makes a difference only if the pattern
931        starts with a lookbehind assertion or \b or \B (\K is not supported by
932        pcre_dfa_exec()). It's an incompatible change, but it makes the two
933        matching functions compatible, and I think it's the right thing to do.
934    
935    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
936        so that the demonstration program is easily available in environments where
937        PCRE has not been installed from source.
938    
939    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
940        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
941        library.
942    
943    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
944        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
945        is not the first non-POSIX option to be added. Clearly some people find
946        these options useful.
947    
948    17. If a caller to the POSIX matching function regexec() passes a non-zero
949        value for nmatch with a NULL value for pmatch, the value of
950        nmatch is forced to zero.
951    
952    18. RunGrepTest did not have a test for the availability of the -u option of
953        the diff command, as RunTest does. It now checks in the same way as
954        RunTest, and also checks for the -b option.
955    
956    19. If an odd number of negated classes containing just a single character
957        interposed, within parentheses, between a forward reference to a named
958        subpattern and the definition of the subpattern, compilation crashed with
959        an internal error, complaining that it could not find the referenced
960        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
961        [The bug was that it was starting one character too far in when skipping
962        over the character class, thus treating the ] as data rather than
963        terminating the class. This meant it could skip too much.]
964    
965    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
966        /g option in pcretest when the pattern contains \K, which makes it possible
967        to have an empty string match not at the start, even when the pattern is
968        anchored. Updated pcretest and pcredemo to use this option.
969    
970    21. If the maximum number of capturing subpatterns in a recursion was greater
971        than the maximum at the outer level, the higher number was returned, but
972        with unset values at the outer level. The correct (outer level) value is
973        now given.
974    
975    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
976        PCRE did not set those parentheses (unlike Perl). I have now found a way to
977        make it do so. The string so far is captured, making this feature
978        compatible with Perl.
979    
980    23. The tests have been re-organized, adding tests 11 and 12, to make it
981        possible to check the Perl 5.10 features against Perl 5.10.
982    
983    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
984        pattern matches a fixed length string. PCRE did not allow this; now it
985        does. Neither allows recursion.
986    
987    25. I finally figured out how to implement a request to provide the minimum
988        length of subject string that was needed in order to match a given pattern.
989        (It was back references and recursion that I had previously got hung up
990        on.) This code has now been added to pcre_study(); it finds a lower bound
991        to the length of subject needed. It is not necessarily the greatest lower
992        bound, but using it to avoid searching strings that are too short does give
993        some useful speed-ups. The value is available to calling programs via
994        pcre_fullinfo().
995    
996    26. While implementing 25, I discovered to my embarrassment that pcretest had
997        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
998        study optimizations had never been tested with that matching function.
999        Oops. What is worse, even when it was passed study data, there was a bug in
1000        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1001        were also very few tests of studied patterns with pcre_dfa_exec().
1002    
1003    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1004        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1005        on the other side of the coin, they are no longer allowed to have different
1006        names, because these cannot be distinguished in PCRE, and this has caused
1007        confusion. (This is a difference from Perl.)
1008    
1009    28. When duplicate subpattern names are present (necessarily with different
1010        numbers, as required by 27 above), and a test is made by name in a
1011        conditional pattern, either for a subpattern having been matched, or for
1012        recursion in such a pattern, all the associated numbered subpatterns are
1013        tested, and the overall condition is true if the condition is true for any
1014        one of them. This is the way Perl works, and is also more like the way
1015        testing by number works.
1016    
1017    
1018    Version 7.9 11-Apr-09
1019    ---------------------
1020    
1021    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1022        (pcretest), all targets were linked against these libraries. This included
1023        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1024        libraries. This caused unwanted dependencies to be created. This problem
1025        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1026        pcretest is linked with readline.
1027    
1028    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1029        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1030        moved outside it again, because FALSE and TRUE are already defined in AIX,
1031        but BOOL is not.
1032    
1033    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1034        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1035    
1036    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1037        hyphens) following file names and line numbers when outputting matching
1038        lines. This is not true; no spaces are inserted. I have also clarified the
1039        wording for the --colour (or --color) option.
1040    
1041    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1042        was not coloured; this is different to GNU grep, so I have changed it to be
1043        the same.
1044    
1045    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1046        each matching line was coloured. Now it goes on to look for further matches
1047        of any of the test patterns, which is the same behaviour as GNU grep.
1048    
1049    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1050        doesn't make sense to accept an empty string match in pcregrep, so I have
1051        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1052        seems to be how GNU grep behaves.
1053    
1054    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1055        start or after a newline", because the conditional assertion was not being
1056        correctly handled. The rule now is that both the assertion and what follows
1057        in the first alternative must satisfy the test.
1058    
1059    9.  If auto-callout was enabled in a pattern with a conditional group whose
1060        condition was an assertion, PCRE could crash during matching, both with
1061        pcre_exec() and pcre_dfa_exec().
1062    
1063    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1064        used for matching.
1065    
1066    11. Unicode property support in character classes was not working for
1067        characters (bytes) greater than 127 when not in UTF-8 mode.
1068    
1069    12. Added the -M command line option to pcretest.
1070    
1071    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1072    
1073    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1074    
1075    16. Added comments and documentation about mis-use of no_arg in the C++
1076        wrapper.
1077    
1078    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1079        from Martin Jerabek that uses macro names for all relevant character and
1080        string constants.
1081    
1082    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1083        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1084        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1085        these, but not everybody uses configure.
1086    
1087    19. A conditional group that had only one branch was not being correctly
1088        recognized as an item that could match an empty string. This meant that an
1089        enclosing group might also not be so recognized, causing infinite looping
1090        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1091        with the subject "ab", where knowledge that the repeated group can match
1092        nothing is needed in order to break the loop.
1093    
1094    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1095        exec(), but without supplying a callout function, matching went wrong.
1096    
1097    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1098        leak if the size of the offset vector was greater than 30. When the vector
1099        is smaller, the saved offsets during recursion go onto a local stack
1100        vector, but for larger vectors malloc() is used. It was failing to free
1101        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1102        error, in fact).
1103    
1104    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1105        heapframe that is used only when UTF-8 support is enabled. This caused no
1106        problem, but was untidy.
1107    
1108    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1109        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1110        included within another project.
1111    
1112    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1113        slightly modified by me:
1114    
1115          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1116              not building pcregrep.
1117    
1118          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1119              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1120    
1121    25. Forward references, both numeric and by name, in patterns that made use of
1122        duplicate group numbers, could behave incorrectly or give incorrect errors,
1123        because when scanning forward to find the reference group, PCRE was not
1124        taking into account the duplicate group numbers. A pattern such as
1125        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1126    
1127    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1128        the feature of a custom pointer more persuasive (as requested by a user).
1129    
1130    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1131        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1132        pre-defined.
1133    
1134    28. Added support for (*UTF8) at the start of a pattern.
1135    
1136    29. Arrange for flags added by the "release type" setting in CMake to be shown
1137        in the configuration summary.
1138    
1139    
1140    Version 7.8 05-Sep-08
1141    ---------------------
1142    
1143    1.  Replaced UCP searching code with optimized version as implemented for Ad
1144        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1145        stage table and inline lookup instead of a function, giving speed ups of 2
1146        to 5 times on some simple patterns that I tested. Permission was given to
1147        distribute the MultiStage2.py script that generates the tables (it's not in
1148        the tarball, but is in the Subversion repository).
1149    
1150    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1151        scripts.
1152    
1153    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1154        a group with a zero qualifier. The result of the study could be incorrect,
1155        or the function might crash, depending on the pattern.
1156    
1157    4.  Caseless matching was not working for non-ASCII characters in back
1158        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1159        It now works when Unicode Property Support is available.
1160    
1161    5.  In pcretest, an escape such as \x{de} in the data was always generating
1162        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1163        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1164        truncation.
1165    
1166    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1167    
1168    7.  Added two (int) casts to pcregrep when printing the difference of two
1169        pointers, in case they are 64-bit values.
1170    
1171    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1172        test 2 if it fails.
1173    
1174    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1175        and a #define of that name to empty if it is not externally set. This is to
1176        allow users of MSVC to set it if necessary.
1177    
1178    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1179        the convenience functions in the pcre_get.c source file.
1180    
1181    11. An option change at the start of a pattern that had top-level alternatives
1182        could cause overwriting and/or a crash. This command provoked a crash in
1183        some environments:
1184    
1185          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1186    
1187        This potential security problem was recorded as CVE-2008-2371.
1188    
1189    12. For a pattern where the match had to start at the beginning or immediately
1190        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1191        pcre_dfa_exec() could read past the end of the passed subject if there was
1192        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1193        pcretest so that it places the subject at the end of its malloc-ed buffer.
1194    
1195    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1196        exec() might read past the end of the data buffer in UTF-8 mode.
1197    
1198    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1199        the data contained the byte 0x85 as part of a UTF-8 character within its
1200        first line. This applied both to normal and DFA matching.
1201    
1202    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1203        /^[^d]*?$/8 failed to match "abc".
1204    
1205    16. Added a missing copyright notice to pcrecpp_internal.h.
1206    
1207    17. Make it more clear in the documentation that values returned from
1208        pcre_exec() in ovector are byte offsets, not character counts.
1209    
1210    18. Tidied a few places to stop certain compilers from issuing warnings.
1211    
1212    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1213        supplied by Stefan Weber. I made a further small update for 7.8 because
1214        there is a change of source arrangements: the pcre_searchfuncs.c module is
1215        replaced by pcre_ucd.c.
1216    
1217    
1218    Version 7.7 07-May-08
1219    ---------------------
1220    
1221    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1222        a string to a long long, pretend we don't even have a long long." This is
1223        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1224    
1225    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1226        pre-7.6 versions, which defined a global no_arg variable instead of putting
1227        it in the RE class. (See also #8 below.)
1228    
1229    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1230        Lopes.
1231    
1232    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1233    
1234        (1) The include/exclude patterns were being applied to the whole pathnames
1235            of files, instead of just to the final components.
1236    
1237        (2) If there was more than one level of directory, the subdirectories were
1238            skipped unless they satisfied the include/exclude conditions. This is
1239            inconsistent with GNU grep (and could even be seen as contrary to the
1240            pcregrep specification - which I improved to make it absolutely clear).
1241            The action now is always to scan all levels of directory, and just
1242            apply the include/exclude patterns to regular files.
1243    
1244    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1245        --exclude_dir in the tests to avoid scanning .svn directories.
1246    
1247    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1248        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1249        doesn't support NULs in patterns.
1250    
1251    7.  Added some missing "const"s to declarations of static tables in
1252        pcre_compile.c and pcre_dfa_exec.c.
1253    
1254    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1255        caused by fix #2  above. (Subsequently also a second patch to fix the
1256        first patch. And a third patch - this was a messy problem.)
1257    
1258    9.  Applied Craig's patch to remove the use of push_back().
1259    
1260    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1261        matching function regexec().
1262    
1263    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1264        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1265        references. PCRE supports relative numbers with this syntax (I don't think
1266        Oniguruma does).
1267    
1268    12. Previously, a group with a zero repeat such as (...){0} was completely
1269        omitted from the compiled regex. However, this means that if the group
1270        was called as a subroutine from elsewhere in the pattern, things went wrong
1271        (an internal error was given). Such groups are now left in the compiled
1272        pattern, with a new opcode that causes them to be skipped at execution
1273        time.
1274    
1275    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1276        to the way PCRE behaves:
1277    
1278        (a) A lone ] character is dis-allowed (Perl treats it as data).
1279    
1280        (b) A back reference to an unmatched subpattern matches an empty string
1281            (Perl fails the current match path).
1282    
1283        (c) A data ] in a character class must be notated as \] because if the
1284            first data character in a class is ], it defines an empty class. (In
1285            Perl it is not possible to have an empty class.) The empty class []
1286            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1287            The negative empty class [^] matches any one character, independently
1288            of the DOTALL setting.
1289    
1290    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1291        non-existent subpattern following a character class starting with ']' and
1292        containing () gave an internal compiling error instead of "reference to
1293        non-existent subpattern". Fortunately, when the pattern did exist, the
1294        compiled code was correct. (When scanning forwards to check for the
1295        existencd of the subpattern, it was treating the data ']' as terminating
1296        the class, so got the count wrong. When actually compiling, the reference
1297        was subsequently set up correctly.)
1298    
1299    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1300        it was being rejected as not supported by pcre_dfa_exec(), even though
1301        other assertions are supported. I have made pcre_dfa_exec() support
1302        (*FAIL).
1303    
1304    16. The implementation of 13c above involved the invention of a new opcode,
1305        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1306        cannot be changed at match time, I realized I could make a small
1307        improvement to matching performance by compiling OP_ALLANY instead of
1308        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1309        on the OP_ANY path.
1310    
1311    17. Compiling pcretest on Windows with readline support failed without the
1312        following two fixes: (1) Make the unistd.h include conditional on
1313        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1314    
1315    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1316        ncurses library to be included for pcretest when ReadLine support is
1317        requested, but also to allow for it to be overridden. This patch came from
1318        Daniel Bergström.
1319    
1320    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1321        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1322        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1323        spotting this.
1324    
1325    
1326    Version 7.6 28-Jan-08
1327    ---------------------
1328    
1329    1.  A character class containing a very large number of characters with
1330        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1331        overflow.
1332    
1333    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1334        HAVE_LONG_LONG is not defined.
1335    
1336    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1337        bring them up to date and include new features. This patch includes:
1338    
1339        - Fixed PH's badly added libz and libbz2 support.
1340        - Fixed a problem with static linking.
1341        - Added pcredemo. [But later removed - see 7 below.]
1342        - Fixed dftables problem and added an option.
1343        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1344            HAVE_LONG_LONG.
1345        - Added readline support for pcretest.
1346        - Added an listing of the option settings after cmake has run.
1347    
1348    4.  A user submitted a patch to Makefile that makes it easy to create
1349        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1350        Makefile.am that cause it to include this special target, without
1351        affecting anything else. Note that the same mingw target plus all
1352        the other distribution libraries and programs are now supported
1353        when configuring with CMake (see 6 below) instead of with
1354        Configure/Make.
1355    
1356    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1357        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1358        exported in the Windows port". It has not yet been confirmed that the patch
1359        solves the problem, but it does no harm.
1360    
1361    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1362        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1363        with CMake, and also correct the comment about stack recursion.
1364    
1365    7.  Remove the automatic building of pcredemo from the ./configure system and
1366        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1367        of a program that users should build themselves after PCRE is installed, so
1368        building it automatically is not really right. What is more, it gave
1369        trouble in some build environments.
1370    
1371    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1372    
1373    
1374    Version 7.5 10-Jan-08
1375    ---------------------
1376    
1377    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1378        values in parens when parsing an RE using the C++ wrapper."
1379    
1380    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1381        Characters greater than 255 were excluded from the class instead of being
1382        included.
1383    
1384    3.  The same bug as (2) above applied to negated POSIX classes such as
1385        [:^space:].
1386    
1387    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1388        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1389        I have changed it.
1390    
1391    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1392        first named subpattern) and a construct such as (?&a) would reference the
1393        first named subpattern whose name started with "a" (in other words, the
1394        length check was missing). Both these problems are fixed. "Subpattern name
1395        expected" is now given for (?&) (a zero-length name), and this patch also
1396        makes it give the same error for \k'' (previously it complained that that
1397        was a reference to a non-existent subpattern).
1398    
1399    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1400        this is right because (?- can be followed by option settings as well as by
1401        digits. I have, however, made the messages clearer.
1402    
1403    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1404        than the number used in the conditional) now cause a compile-time error.
1405        This is actually not compatible with Perl, which accepts such patterns, but
1406        treats the conditional as always being FALSE (as PCRE used to), but it
1407        seems to me that giving a diagnostic is better.
1408    
1409    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1410        and messages.
1411    
1412    9.  Fix two occurrences of "backslash" in comments that should have been
1413        "backspace".
1414    
1415    10. Remove two redundant lines of code that can never be obeyed (their function
1416        was moved elsewhere).
1417    
1418    11. The program that makes PCRE's Unicode character property table had a bug
1419        which caused it to generate incorrect table entries for sequences of
1420        characters that have the same character type, but are in different scripts.
1421        It amalgamated them into a single range, with the script of the first of
1422        them. In other words, some characters were in the wrong script. There were
1423        thirteen such cases, affecting characters in the following ranges:
1424    
1425          U+002b0 - U+002c1
1426          U+0060c - U+0060d
1427          U+0061e - U+00612
1428          U+0064b - U+0065e
1429          U+0074d - U+0076d
1430          U+01800 - U+01805
1431          U+01d00 - U+01d77
1432          U+01d9b - U+01dbf
1433          U+0200b - U+0200f
1434          U+030fc - U+030fe
1435          U+03260 - U+0327f
1436          U+0fb46 - U+0fbb1
1437          U+10450 - U+1049d
1438    
1439    12. The -o option (show only the matching part of a line) for pcregrep was not
1440        compatible with GNU grep in that, if there was more than one match in a
1441        line, it showed only the first of them. It now behaves in the same way as
1442        GNU grep.
1443    
1444    13. If the -o and -v options were combined for pcregrep, it printed a blank
1445        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1446        does the same. The return code can be used to tell if there were any
1447        non-matching lines.
1448    
1449    14. Added --file-offsets and --line-offsets to pcregrep.
1450    
1451    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1452        infinitely looping recursion. The bug was that positive lookaheads were not
1453        being skipped when checking for a possible empty match (negative lookaheads
1454        and both kinds of lookbehind were skipped).
1455    
1456    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1457        inclusion of <windows.h> to before rather than after the definition of
1458        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1459    
1460    17. Specifying a possessive quantifier with a specific limit for a Unicode
1461        character property caused pcre_compile() to compile bad code, which led at
1462        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1463        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1464        caused the error; without that there was no problem.
1465    
1466    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1467    
1468    19. Added --enable-pcretest-libreadline.
1469    
1470    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1471        RE::GlobalReplace(). As a result, the number of replacements returned was
1472        double what it should be. I removed one of the increments, but Craig sent a
1473        later patch that removed the other one (the right fix) and added unit tests
1474        that check the return values (which was not done before).
1475    
1476    21. Several CMake things:
1477    
1478        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1479            the names libpcre and libpcreposix, not just pcre and pcreposix.
1480    
1481        (2) The above change means that pcretest and pcregrep are now correctly
1482            linked with the newly-built libraries, not previously installed ones.
1483    
1484        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1485    
1486    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1487        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1488        UTF-8 newline character). The key issue is that the pattern starts .*;
1489        this means that the match must be either at the beginning, or after a
1490        newline. The bug was in the code for advancing after a failed match and
1491        checking that the new position followed a newline. It was not taking
1492        account of UTF-8 characters correctly.
1493    
1494    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1495        character classes. PCRE was not treating the sequence [:...:] as a
1496        character class unless the ... were all letters. Perl, however, seems to
1497        allow any characters between [: and :], though of course it rejects as
1498        unknown any "names" that contain non-letters, because all the known class
1499        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1500        for example, whereas PCRE did not - it did not recognize a POSIX character
1501        class. This seemed a bit dangerous, so the code has been changed to be
1502        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1503        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1504        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1505        Perl does, and where it didn't before.
1506    
1507    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1508        Windows environments %n is disabled by default.
1509    
1510    
1511    Version 7.4 21-Sep-07
1512    ---------------------
1513    
1514    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1515        means that a class such as [\s] counted as "explicit reference to CR or
1516        LF". That isn't really right - the whole point of the change was to try to
1517        help when there was an actual mention of one of the two characters. So now
1518        the change happens only if \r or \n (or a literal CR or LF) character is
1519        encountered.
1520    
1521    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1522        of both had grown to the point where there were only 3 bits left.
1523        Fortunately, there was spare space in the data structure, and so I have
1524        moved the internal flags into a new 16-bit field to free up more option
1525        bits.
1526    
1527    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1528        but did not set the internal JCHANGED flag - either of these is enough to
1529        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1530        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1531        start sets both bits.
1532    
1533    4.  Added options (at build time, compile time, exec time) to change \R from
1534        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1535    
1536    5.  doc/pcresyntax.html was missing from the distribution.
1537    
1538    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1539        compatibility, even though it is no longer used.
1540    
1541    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1542        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1543        windows.h file is present (where different names are used). [This was
1544        reversed later after testing - see 16 below.]
1545    
1546    8.  Changed all #include <config.h> to #include "config.h". There were also
1547        some further <pcre.h> cases that I changed to "pcre.h".
1548    
1549    9.  When pcregrep was used with the --colour option, it missed the line ending
1550        sequence off the lines that it output.
1551    
1552    10. It was pointed out to me that arrays of string pointers cause lots of
1553        relocations when a shared library is dynamically loaded. A technique of
1554        using a single long string with a table of offsets can drastically reduce
1555        these. I have refactored PCRE in four places to do this. The result is
1556        dramatic:
1557    
1558          Originally:                          290
1559          After changing UCP table:            187
1560          After changing error message table:   43
1561          After changing table of "verbs"       36
1562          After changing table of Posix names   22
1563    
1564        Thanks to the folks working on Gregex for glib for this insight.
1565    
1566    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1567        unicode-properties was also set.
1568    
1569    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1570    
1571    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1572        checked only for CRLF.
1573    
1574    14. Added casts to pcretest.c to avoid compiler warnings.
1575    
1576    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1577    
1578    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1579        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1580        entirely. This removes changes made in 7 above.
1581    
1582    17. The CMake files have been updated, and there is now more information about
1583        building with CMake in the NON-UNIX-USE document.
1584    
1585    
1586    Version 7.3 28-Aug-07
1587  ---------------------  ---------------------
1588    
1589   1. In the rejigging of the build system that eventually resulted in 7.1, the   1. In the rejigging of the build system that eventually resulted in 7.1, the
# Line 98  Version 7.3 09-Aug-07 Line 1680  Version 7.3 09-Aug-07
1680      the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the      the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1681      full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still      full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1682      does: it's just the validity check that is more restrictive.      does: it's just the validity check that is more restrictive.
1683    
1684  16. Inserted checks for integer overflows during escape sequence (backslash)  16. Inserted checks for integer overflows during escape sequence (backslash)
1685      processing, and also fixed erroneous offset values for syntax errors during      processing, and also fixed erroneous offset values for syntax errors during
1686      backslash processing.      backslash processing.
1687    
1688  17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)  17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1689      for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".      for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1690    
1691  18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"  18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1692      caused an overrun.      caused an overrun.
1693    
1694  19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with  19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1695      something other than just ASCII characters) inside a group that had an      something other than just ASCII characters) inside a group that had an
1696      unlimited repeat caused a loop at compile time (while checking to see      unlimited repeat caused a loop at compile time (while checking to see
1697      whether the group could match an empty string).      whether the group could match an empty string).
1698    
1699  20. Debugging a pattern containing \p or \P could cause a crash. For example,  20. Debugging a pattern containing \p or \P could cause a crash. For example,
1700      [\P{Any}] did so. (Error in the code for printing property names.)      [\P{Any}] did so. (Error in the code for printing property names.)
   
 21. An orphan \E inside a character class could cause a crash.  
1701    
1702  22. A repeated capturing bracket such as (A)? could cause a wild memory  21. An orphan \E inside a character class could cause a crash.
1703    
1704    22. A repeated capturing bracket such as (A)? could cause a wild memory
1705      reference during compilation.      reference during compilation.
1706    
1707  23. There are several functions in pcre_compile() that scan along a compiled  23. There are several functions in pcre_compile() that scan along a compiled
1708      expression for various reasons (e.g. to see if it's fixed length for look      expression for various reasons (e.g. to see if it's fixed length for look
1709      behind). There were bugs in these functions when a repeated \p or \P was      behind). There were bugs in these functions when a repeated \p or \P was
1710      present in the pattern. These operators have additional parameters compared      present in the pattern. These operators have additional parameters compared
1711      with \d, etc, and these were not being taken into account when moving along      with \d, etc, and these were not being taken into account when moving along
1712      the compiled data. Specifically:      the compiled data. Specifically:
1713    
1714      (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed      (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1715          length.          length.
1716    
1717      (b) An item such as \pL+ within a repeated group could cause crashes or      (b) An item such as \pL+ within a repeated group could cause crashes or
1718          loops.          loops.
1719    
1720      (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect      (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1721          "reference to non-existent subpattern" error.          "reference to non-existent subpattern" error.
1722    
1723        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1724    
1725    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1726        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1727    
1728    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1729    
1730    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1731        character were causing crashes (broken optimization).
1732    
1733    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1734        \p or \P) caused a compile-time loop.
1735    
1736    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1737        break. For example, the unstudied pattern [\r\n]A does not match the string
1738        "\r\nA" because change 7.0/46 below moves the current point on by two
1739        characters after failing to match at the start. However, the pattern \nA
1740        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1741        the same is true. There doesn't seem any very clean way out of this, but
1742        what I have chosen to do makes the common cases work: PCRE now takes note
1743        of whether there can be an explicit match for \r or \n anywhere in the
1744        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1745        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1746        pattern has explicit CR or LF references.
1747    
1748    29. Added (*CR) etc for changing newline setting at start of pattern.
1749    
1750    
1751  Version 7.2 19-Jun-07  Version 7.2 19-Jun-07

Legend:
Removed from v.218  
changed lines
  Added in v.747

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12