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

Diff of /code/trunk/ChangeLog

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

revision 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 758 by ph10, Mon Nov 21 12:05:36 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.21
5    ------------
6    
7    1.  Updating the JIT compiler.
8    
9    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10        are added as well.
11    
12    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14        calling _pcre_jit_exec. Some extra comments are added.
15    
16    4.  Mark settings inside atomic groups that do not contain any capturing
17        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18        was introduced by change 18 for 8.20.
19    
20    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21        ECMA-262 standard.
22    
23    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25        This bug was probably introduced by change 9 of 8.13.
26    
27    7.  While fixing 6 above, I noticed that a number of other items were being
28        incorrectly rejected as "not fixed length". This arose partly because newer
29        opcodes had not been added to the fixed-length checking code. I have (a)
30        corrected the bug and added tests for these items, and (b) arranged for an
31        error to occur if an unknown opcode is encountered while checking for fixed
32        length instead of just assuming "not fixed length". The items that were
33        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36    
37    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
38        being incorrectly compiled and would have given unpredicatble results.
39    
40    9.  A possessively repeated subpattern with minimum repeat count greater than
41        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
42        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
43        the first (A) could occur when it should not.
44    
45    10. Add a cast and remove a redundant test from the code.
46    
47    11. JIT should use pcre_malloc/pcre_free for allocation.
48    
49    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
50        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
51        is anything other than /usr, -L is still shown).
52    
53    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
54    
55    14. Perl does not support \N without a following name in a [] class; PCRE now
56        also gives an error.
57    
58    
59    Version 8.20 21-Oct-2011
60    ------------------------
61    
62    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
63        a POSIX class. After further experiments with Perl, which convinced me that
64        Perl has bugs and confusions, a closing square bracket is no longer allowed
65        in a POSIX name. This bug also affected patterns with classes that started
66        with full stops.
67    
68    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
69        captured substring, but while checking the failing first alternative,
70        substring 1 is temporarily captured. If the output vector supplied to
71        pcre_exec() was not big enough for this capture, the yield of the function
72        was still zero ("insufficient space for captured substrings"). This cannot
73        be totally fixed without adding another stack variable, which seems a lot
74        of expense for a edge case. However, I have improved the situation in cases
75        such as /(a)(b)x|abc/ matched against "abc", where the return code
76        indicates that fewer than the maximum number of slots in the ovector have
77        been set.
78    
79    3.  Related to (2) above: when there are more back references in a pattern than
80        slots in the output vector, pcre_exec() uses temporary memory during
81        matching, and copies in the captures as far as possible afterwards. It was
82        using the entire output vector, but this conflicts with the specification
83        that only 2/3 is used for passing back captured substrings. Now it uses
84        only the first 2/3, for compatibility. This is, of course, another edge
85        case.
86    
87    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
88        main code base, and can be used by building with --enable-jit. When this is
89        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
90        runtime --no-jit option is given.
91    
92    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
93        ovector, the return from the function was zero, implying that there were
94        other matches that did not fit. The correct "exactly full" value is now
95        returned.
96    
97    6.  If a subpattern that was called recursively or as a subroutine contained
98        (*PRUNE) or any other control that caused it to give a non-standard return,
99        invalid errors such as "Error -26 (nested recursion at the same subject
100        position)" or even infinite loops could occur.
101    
102    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
103        computing the minimum length on reaching *ACCEPT, and so ended up with the
104        wrong value of 1 rather than 0. Further investigation indicates that
105        computing a minimum subject length in the presence of *ACCEPT is difficult
106        (think back references, subroutine calls), and so I have changed the code
107        so that no minimum is registered for a pattern that contains *ACCEPT.
108    
109    8.  If (*THEN) was present in the first (true) branch of a conditional group,
110        it was not handled as intended. [But see 16 below.]
111    
112    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
113        Sheri Pierce.
114    
115    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
116        the first byte in a match must be "a".
117    
118    11. Change 17 for 8.13 increased the recursion depth for patterns like
119        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
120        pattern contains any instances of (*THEN). If it does not, the old
121        optimizations are restored. It would be nice to do this on a per-group
122        basis, but at the moment that is not feasible.
123    
124    12. In some environments, the output of pcretest -C is CRLF terminated. This
125        broke RunTest's code that checks for the link size. A single white space
126        character after the value is now allowed for.
127    
128    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
129        For "fr", it uses the Windows-specific input and output files.
130    
131    14. If (*THEN) appeared in a group that was called recursively or as a
132        subroutine, it did not work as intended. [But see next item.]
133    
134    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
135        pattern fragments (but not containing any | characters). If A and B are
136        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
137        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
138        D. In other words, Perl considers parentheses that do not contain any |
139        characters to be part of a surrounding alternative, whereas PCRE was
140        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
141        differently. PCRE now behaves in the same way as Perl, except in the case
142        of subroutine/recursion calls such as (?1) which have in any case always
143        been different (but PCRE had them first :-).
144    
145    16. Related to 15 above: Perl does not treat the | in a conditional group as
146        creating alternatives. Such a group is treated in the same way as an
147        ordinary group without any | characters when processing (*THEN). PCRE has
148        been changed to match Perl's behaviour.
149    
150    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
151        RunGrepTest script failed.
152    
153    18. Change 22 for version 13 caused atomic groups to use more stack. This is
154        inevitable for groups that contain captures, but it can lead to a lot of
155        stack use in large patterns. The old behaviour has been restored for atomic
156        groups that do not contain any capturing parentheses.
157    
158    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
159        suppress the check for a minimum subject length at run time. (If it was
160        given to pcre_exec() or pcre_dfa_exec() it did work.)
161    
162    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
163        fail to work when decoding hex characters in data strings in EBCDIC
164        environments.
165    
166    21. It appears that in at least one Mac OS environment, the isxdigit() function
167        is implemented as a macro that evaluates to its argument more than once,
168        contravening the C 90 Standard (I haven't checked a later standard). There
169        was an instance in pcretest which caused it to go wrong when processing
170        \x{...} escapes in subject strings. The has been rewritten to avoid using
171        things like p++ in the argument of isxdigit().
172    
173    
174    Version 8.13 16-Aug-2011
175    ------------------------
176    
177    1.  The Unicode data tables have been updated to Unicode 6.0.0.
178    
179    2.  Two minor typos in pcre_internal.h have been fixed.
180    
181    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
182        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
183        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
184    
185    4.  There were a number of related bugs in the code for matching backrefences
186        caselessly in UTF-8 mode when codes for the characters concerned were
187        different numbers of bytes. For example, U+023A and U+2C65 are an upper
188        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
189        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
190        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
191        2-byte code at the end of the subject (it thought there wasn't enough data
192        left).
193    
194    5.  Comprehensive information about what went wrong is now returned by
195        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
196        as the output vector has at least 2 elements. The offset of the start of
197        the failing character and a reason code are placed in the vector.
198    
199    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
200        now returned is for the first byte of the failing character, instead of the
201        last byte inspected. This is an incompatible change, but I hope it is small
202        enough not to be a problem. It makes the returned offset consistent with
203        pcre_exec() and pcre_dfa_exec().
204    
205    7.  pcretest now gives a text phrase as well as the error number when
206        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
207        failure, the offset and reason code are output.
208    
209    8.  When \R was used with a maximizing quantifier it failed to skip backwards
210        over a \r\n pair if the subsequent match failed. Instead, it just skipped
211        back over a single character (\n). This seems wrong (because it treated the
212        two characters as a single entity when going forwards), conflicts with the
213        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
214        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
215        has been changed.
216    
217    9.  Some internal refactoring has changed the processing so that the handling
218        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
219        time (the PCRE_DOTALL option was changed this way some time ago: version
220        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
221        which was always a bit of a fudge. It also means that there is one less
222        argument for the match() function, which reduces its stack requirements
223        slightly. This change also fixes an incompatibility with Perl: the pattern
224        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
225    
226    10. More internal refactoring has drastically reduced the number of recursive
227        calls to match() for possessively repeated groups such as (abc)++ when
228        using pcre_exec().
229    
230    11. While implementing 10, a number of bugs in the handling of groups were
231        discovered and fixed:
232    
233        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
234        (a|)*(?1) gave a compile-time internal error.
235        ((a|)+)+  did not notice that the outer group could match an empty string.
236        (^a|^)+   was not marked as anchored.
237        (.*a|.*)+ was not marked as matching at start or after a newline.
238    
239    12. Yet more internal refactoring has removed another argument from the match()
240        function. Special calls to this function are now indicated by setting a
241        value in a variable in the "match data" data block.
242    
243    13. Be more explicit in pcre_study() instead of relying on "default" for
244        opcodes that mean there is no starting character; this means that when new
245        ones are added and accidentally left out of pcre_study(), testing should
246        pick them up.
247    
248    14. The -s option of pcretest has been documented for ages as being an old
249        synonym of -m (show memory usage). I have changed it to mean "force study
250        for every regex", that is, assume /S for every regex. This is similar to -i
251        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
252        using it. It makes it easier to run collections of tests with and without
253        study enabled, and thereby test pcre_study() more easily. All the standard
254        tests are now run with and without -s (but some patterns can be marked as
255        "never study" - see 20 below).
256    
257    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
258        restoration of the capturing data to the outer values was not happening
259        correctly.
260    
261    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
262        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
263        pattern had matched an empty string, and so incorrectly returned a no
264        match.
265    
266    17. There was optimizing code for the last branch of non-capturing parentheses,
267        and also for the obeyed branch of a conditional subexpression, which used
268        tail recursion to cut down on stack usage. Unfortunately, now that there is
269        the possibility of (*THEN) occurring in these branches, tail recursion is
270        no longer possible because the return has to be checked for (*THEN). These
271        two optimizations have therefore been removed. [But see 8.20/11 above.]
272    
273    18. If a pattern containing \R was studied, it was assumed that \R always
274        matched two bytes, thus causing the minimum subject length to be
275        incorrectly computed because \R can also match just one byte.
276    
277    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
278        was incorrectly computed.
279    
280    20. If /S is present twice on a test pattern in pcretest input, it now
281        *disables* studying, thereby overriding the use of -s on the command line
282        (see 14 above). This is necessary for one or two tests to keep the output
283        identical in both cases.
284    
285    21. When (*ACCEPT) was used in an assertion that matched an empty string and
286        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
287    
288    22. When an atomic group that contained a capturing parenthesis was
289        successfully matched, but the branch in which it appeared failed, the
290        capturing was not being forgotten if a higher numbered group was later
291        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
292        group 1 to "a", when in fact it should be unset. This applied to multi-
293        branched capturing and non-capturing groups, repeated or not, and also to
294        positive assertions (capturing in negative assertions does not happen
295        in PCRE) and also to nested atomic groups.
296    
297    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
298        subject after a captured substring, to make it easier to tell which of a
299        number of identical substrings has been captured.
300    
301    24. The way atomic groups are processed by pcre_exec() has been changed so that
302        if they are repeated, backtracking one repetition now resets captured
303        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
304        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
305        "aaa". Previously, it would have been "a". As part of this code
306        refactoring, the way recursive calls are handled has also been changed.
307    
308    25. If an assertion condition captured any substrings, they were not passed
309        back unless some other capturing happened later. For example, if
310        (?(?=(a))a) was matched against "a", no capturing was returned.
311    
312    26. When studying a pattern that contained subroutine calls or assertions,
313        the code for finding the minimum length of a possible match was handling
314        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
315        group 1 called group 2 while simultaneously a separate group 2 called group
316        1). A stack overflow occurred in this case. I have fixed this by limiting
317        the recursion depth to 10.
318    
319    27. Updated RunTest.bat in the distribution to the version supplied by Tom
320        Fortmann. This supports explicit test numbers on the command line, and has
321        argument validation and error reporting.
322    
323    28. An instance of \X with an unlimited repeat could fail if at any point the
324        first character it looked at was a mark character.
325    
326    29. Some minor code refactoring concerning Unicode properties and scripts
327        should reduce the stack requirement of match() slightly.
328    
329    30. Added the '=' option to pcretest to check the setting of unused capturing
330        slots at the end of the pattern, which are documented as being -1, but are
331        not included in the return count.
332    
333    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
334        compiled something random. Now it gives a compile-time error (as does
335        Perl).
336    
337    32. A *MARK encountered during the processing of a positive assertion is now
338        recorded and passed back (compatible with Perl).
339    
340    33. If --only-matching or --colour was set on a pcregrep call whose pattern
341        had alternative anchored branches, the search for a second match in a line
342        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
343        matched the line "0102" twice. The same bug affected patterns that started
344        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
345        twice.
346    
347    34. Previously, PCRE did not allow quantification of assertions. However, Perl
348        does, and because of capturing effects, quantifying parenthesized
349        assertions may at times be useful. Quantifiers are now allowed for
350        parenthesized assertions.
351    
352    35. A minor code tidy in pcre_compile() when checking options for \R usage.
353    
354    36. \g was being checked for fancy things in a character class, when it should
355        just be a literal "g".
356    
357    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
358        appearance of a nested POSIX class supersedes an apparent external class.
359        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
360        unescaped square brackets may also appear as part of class names. For
361        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
362        more like Perl. (But see 8.20/1 above.)
363    
364    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
365        was because it thought it was \N{name}, which is not supported).
366    
367    39. Add minix to OS list not supporting the -S option in pcretest.
368    
369    40. PCRE tries to detect cases of infinite recursion at compile time, but it
370        cannot analyze patterns in sufficient detail to catch mutual recursions
371        such as ((?1))((?2)). There is now a runtime test that gives an error if a
372        subgroup is called recursively as a subpattern for a second time at the
373        same position in the subject string. In previous releases this might have
374        been caught by the recursion limit, or it might have run out of stack.
375    
376    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
377        happen only once. PCRE was, however incorrectly giving a compile time error
378        "recursive call could loop indefinitely" because it cannot analyze the
379        pattern in sufficient detail. The compile time test no longer happens when
380        PCRE is compiling a conditional subpattern, but actual runaway loops are
381        now caught at runtime (see 40 above).
382    
383    42. It seems that Perl allows any characters other than a closing parenthesis
384        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
385        has been changed to be the same.
386    
387    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
388        as not to get warnings when autogen.sh is called. Also changed
389        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
390    
391    44. To help people who use pcregrep to scan files containing exceedingly long
392        lines, the following changes have been made:
393    
394        (a) The default value of the buffer size parameter has been increased from
395            8K to 20K. (The actual buffer used is three times this size.)
396    
397        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
398            PCRE is built.
399    
400        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
401            to be set at run time.
402    
403        (d) Numerical values in pcregrep options can be followed by K or M, for
404            example --buffer-size=50K.
405    
406        (e) If a line being scanned overflows pcregrep's buffer, an error is now
407            given and the return code is set to 2.
408    
409    45. Add a pointer to the latest mark to the callout data block.
410    
411    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
412        partial match of an empty string instead of no match. This was specific to
413        the use of ".".
414    
415    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
416        complete match instead of a partial match. This bug was dependent on both
417        the PCRE_UTF8 and PCRE_DOTALL options being set.
418    
419    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
420        starting byte set, because \b was not being ignored.
421    
422    
423    Version 8.12 15-Jan-2011
424    ------------------------
425    
426    1.  Fixed some typos in the markup of the man pages, and wrote a script that
427        checks for such things as part of the documentation building process.
428    
429    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
430        --match-limit and --recursion-limit options (added for 8.11). In
431        particular, this made one of the standard tests fail. (The integer value
432        went into the wrong half of a long int.)
433    
434    3.  If the --colour option was given to pcregrep with -v (invert match), it
435        did strange things, either producing crazy output, or crashing. It should,
436        of course, ignore a request for colour when reporting lines that do not
437        match.
438    
439    4.  Another pcregrep bug caused similar problems if --colour was specified with
440        -M (multiline) and the pattern match finished with a line ending.
441    
442    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
443        matched in multiline mode, the following line was shown as part of the
444        match. This seems wrong, so I have changed it.
445    
446    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
447        the check for further matches in the same line (so they could be coloured)
448        to overrun the end of the current line. If another match was found, it was
449        incorrectly shown (and then shown again when found in the next line).
450    
451    7.  If pcregrep was compiled under Windows, there was a reference to the
452        function pcregrep_exit() before it was defined. I am assuming this was
453        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
454        reported by a user. I've moved the definition above the reference.
455    
456    
457    Version 8.11 10-Dec-2010
458    ------------------------
459    
460    1.  (*THEN) was not working properly if there were untried alternatives prior
461        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
462        backtracked to try for "b" instead of moving to the next alternative branch
463        at the same level (in this case, to look for "c"). The Perl documentation
464        is clear that when (*THEN) is backtracked onto, it goes to the "next
465        alternative in the innermost enclosing group".
466    
467    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
468        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
469        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
470        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
471        (*THEN).
472    
473    3.  If \s appeared in a character class, it removed the VT character from
474        the class, even if it had been included by some previous item, for example
475        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
476        of \s, but is part of the POSIX "space" class.)
477    
478    4.  A partial match never returns an empty string (because you can always
479        match an empty string at the end of the subject); however the checking for
480        an empty string was starting at the "start of match" point. This has been
481        changed to the "earliest inspected character" point, because the returned
482        data for a partial match starts at this character. This means that, for
483        example, /(?<=abc)def/ gives a partial match for the subject "abc"
484        (previously it gave "no match").
485    
486    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
487        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
488        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
489        has an implication that the given string is incomplete (because a partial
490        match is preferred over a full match). For this reason, these items now
491        give a partial match in this situation. [Aside: previously, the one case
492        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
493        match rather than a full match, which was wrong by the old rules, but is
494        now correct.]
495    
496    6.  There was a bug in the handling of #-introduced comments, recognized when
497        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
498        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
499        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
500        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
501        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
502        places in pcre_compile().
503    
504    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
505        comments when looking ahead for named forward references to subpatterns,
506        the only newline sequence it recognized was NL. It now handles newlines
507        according to the set newline convention.
508    
509    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
510        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
511        cater for a lack of strerror(). These oversights have been fixed.
512    
513    9.  Added --match-limit and --recursion-limit to pcregrep.
514    
515    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
516    
517    11. When the -o option was used, pcregrep was setting a return code of 1, even
518        when matches were found, and --line-buffered was not being honoured.
519    
520    12. Added an optional parentheses number to the -o and --only-matching options
521        of pcregrep.
522    
523    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
524        can match an empty string. The code to do it in pcretest and pcredemo
525        needed fixing:
526    
527        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
528            only one byte after an empty string match just before CRLF (this case
529            just got forgotten; "any" and "anycrlf" were OK).
530    
531        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
532            mode when an empty string match preceded an ASCII character followed by
533            a non-ASCII character. (The code for advancing by one character rather
534            than one byte was nonsense.)
535    
536        (c) The pcredemo.c sample program did not have any code at all to handle
537            the cases when CRLF is a valid newline sequence.
538    
539    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
540        as a starting offset was within the subject string. There is now a new
541        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
542        negative or greater than the length of the string. In order to test this,
543        pcretest is extended to allow the setting of negative starting offsets.
544    
545    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
546        starting offset points to the beginning of a UTF-8 character was
547        unnecessarily clumsy. I tidied it up.
548    
549    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
550        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
551    
552    17. Nobody had reported that the --include_dir option, which was added in
553        release 7.7 should have been called --include-dir (hyphen, not underscore)
554        for compatibility with GNU grep. I have changed it to --include-dir, but
555        left --include_dir as an undocumented synonym, and the same for
556        --exclude-dir, though that is not available in GNU grep, at least as of
557        release 2.5.4.
558    
559    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
560        characters from a string of bytes) have been redefined so as not to use
561        loops, in order to improve performance in some environments. At the same
562        time, I abstracted some of the common code into auxiliary macros to save
563        repetition (this should not affect the compiled code).
564    
565    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
566        compile-time error is now given if \c is not followed by an ASCII
567        character, that is, a byte less than 128. (In EBCDIC mode, the code is
568        different, and any byte value is allowed.)
569    
570    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
571        START_OPTIMIZE option, which is now allowed at compile time - but just
572        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
573        to pcregrep and other applications that have no direct access to PCRE
574        options. The new /Y option in pcretest sets this option when calling
575        pcre_compile().
576    
577    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
578        back references. Groups containing recursive back references were forced to
579        be atomic by that change, but in the case of named groups, the amount of
580        memory required was incorrectly computed, leading to "Failed: internal
581        error: code overflow". This has been fixed.
582    
583    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
584        pcretest.c, to avoid build problems in some Borland environments.
585    
586    
587    Version 8.10 25-Jun-2010
588    ------------------------
589    
590    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
591        THEN.
592    
593    2.  (*ACCEPT) was not working when inside an atomic group.
594    
595    3.  Inside a character class, \B is treated as a literal by default, but
596        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
597        causes the error). The code is unchanged, but I tidied the documentation.
598    
599    4.  Inside a character class, PCRE always treated \R and \X as literals,
600        whereas Perl faults them if its -w option is set. I have changed PCRE so
601        that it faults them when PCRE_EXTRA is set.
602    
603    5.  Added support for \N, which always matches any character other than
604        newline. (It is the same as "." when PCRE_DOTALL is not set.)
605    
606    6.  When compiling pcregrep with newer versions of gcc which may have
607        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
608        declared with attribute warn_unused_result" were given. Just casting the
609        result to (void) does not stop the warnings; a more elaborate fudge is
610        needed. I've used a macro to implement this.
611    
612    7.  Minor change to pcretest.c to avoid a compiler warning.
613    
614    8.  Added four artifical Unicode properties to help with an option to make
615        \s etc use properties (see next item). The new properties are: Xan
616        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
617    
618    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
619        use Unicode properties. (*UCP) at the start of a pattern can be used to set
620        this option. Modified pcretest to add /W to test this facility. Added
621        REG_UCP to make it available via the POSIX interface.
622    
623    10. Added --line-buffered to pcregrep.
624    
625    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
626        studied, and the match started with a letter with a code point greater than
627        127 whose first byte was different to the first byte of the other case of
628        the letter, the other case of this starting letter was not recognized
629        (#976).
630    
631    12. If a pattern that was studied started with a repeated Unicode property
632        test, for example, \p{Nd}+, there was the theoretical possibility of
633        setting up an incorrect bitmap of starting bytes, but fortunately it could
634        not have actually happened in practice until change 8 above was made (it
635        added property types that matched character-matching opcodes).
636    
637    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
638        possible starting bytes for non-anchored patterns.
639    
640    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
641        \R, and also a number of cases that involve Unicode properties, both
642        explicit and implicit when PCRE_UCP is set.
643    
644    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
645        input, it could crash or give wrong results if characters with values
646        greater than 0xc0 were present in the subject string. (Detail: it assumed
647        UTF-8 input when processing these items.)
648    
649    16. Added a lot of (int) casts to avoid compiler warnings in systems where
650        size_t is 64-bit (#991).
651    
652    17. Added a check for running out of memory when PCRE is compiled with
653        --disable-stack-for-recursion (#990).
654    
655    18. If the last data line in a file for pcretest does not have a newline on
656        the end, a newline was missing in the output.
657    
658    19. The default pcre_chartables.c file recognizes only ASCII characters (values
659        less than 128) in its various bitmaps. However, there is a facility for
660        generating tables according to the current locale when PCRE is compiled. It
661        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
662        characters, are recognized by isspace() and therefore were getting set in
663        these tables, and indeed these tables seem to approximate to ISO 8859. This
664        caused a problem in UTF-8 mode when pcre_study() was used to create a list
665        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
666        which of course cannot start UTF-8 characters. I have changed the code so
667        that only real ASCII characters (less than 128) and the correct starting
668        bytes for UTF-8 encodings are set for characters greater than 127 when in
669        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
670        altogether.)
671    
672    20. Added the /T option to pcretest so as to be able to run tests with non-
673        standard character tables, thus making it possible to include the tests
674        used for 19 above in the standard set of tests.
675    
676    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
677        reference to a subpattern the other side of a comment that contains an
678        opening parenthesis caused either an internal compiling error, or a
679        reference to the wrong subpattern.
680    
681    
682    Version 8.02 19-Mar-2010
683    ------------------------
684    
685    1.  The Unicode data tables have been updated to Unicode 5.2.0.
686    
687    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
688        configured.
689    
690    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
691        original author of that file, following a query about its status.
692    
693    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
694        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
695    
696    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
697        quantifier applied to a forward-referencing subroutine call, could compile
698        incorrect code or give the error "internal error: previously-checked
699        referenced subpattern not found".
700    
701    6.  Both MS Visual Studio and Symbian OS have problems with initializing
702        variables to point to external functions. For these systems, therefore,
703        pcre_malloc etc. are now initialized to local functions that call the
704        relevant global functions.
705    
706    7.  There were two entries missing in the vectors called coptable and poptable
707        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
708        I've fixed the data, and added a kludgy way of testing at compile time that
709        the lengths are correct (equal to the number of opcodes).
710    
711    8.  Following on from 7, I added a similar kludge to check the length of the
712        eint vector in pcreposix.c.
713    
714    9.  Error texts for pcre_compile() are held as one long string to avoid too
715        much relocation at load time. To find a text, the string is searched,
716        counting zeros. There was no check for running off the end of the string,
717        which could happen if a new error number was added without updating the
718        string.
719    
720    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
721    
722    11. \K was not working if it appeared in an atomic group or in a group that
723        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
724        \K is "not well defined" if used in an assertion. PCRE now accepts it if
725        the assertion is positive, but not if it is negative.
726    
727    12. Change 11 fortuitously reduced the size of the stack frame used in the
728        "match()" function of pcre_exec.c by one pointer. Forthcoming
729        implementation of support for (*MARK) will need an extra pointer on the
730        stack; I have reserved it now, so that the stack frame size does not
731        decrease.
732    
733    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
734        item in branch that calls a recursion is a subroutine call - as in the
735        second branch in the above example - was incorrectly given the compile-
736        time error "recursive call could loop indefinitely" because pcre_compile()
737        was not correctly checking the subroutine for matching a non-empty string.
738    
739    14. The checks for overrunning compiling workspace could trigger after an
740        overrun had occurred. This is a "should never occur" error, but it can be
741        triggered by pathological patterns such as hundreds of nested parentheses.
742        The checks now trigger 100 bytes before the end of the workspace.
743    
744    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
745    
746    
747    Version 8.01 19-Jan-2010
748    ------------------------
749    
750    1.  If a pattern contained a conditional subpattern with only one branch (in
751        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
752        computed the wrong minimum data length (which is of course zero for such
753        subpatterns). This could cause incorrect "no match" results.
754    
755    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
756        the pattern is reset in the first branch, pcre_compile() failed with
757        "internal error: code overflow at offset...". This happened only when
758        the reset was to the original external option setting. (An optimization
759        abstracts leading options settings into an external setting, which was the
760        cause of this.)
761    
762    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
763        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
764        assertion pattern did not match (meaning that the assertion was true), it
765        was incorrectly treated as false if the SKIP had been reached during the
766        matching. This also applied to assertions used as conditions.
767    
768    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
769        assertion subpattern, including such a pattern used as a condition,
770        unpredictable results occurred, instead of the error return
771        PCRE_ERROR_DFA_UITEM.
772    
773    5.  The C++ GlobalReplace function was not working like Perl for the special
774        situation when an empty string is matched. It now does the fancy magic
775        stuff that is necessary.
776    
777    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
778        removed. (These were left over from very, very early versions of PCRE.)
779    
780    7.  Some cosmetic changes to the code to make life easier when compiling it
781        as part of something else:
782    
783        (a) Change DEBUG to PCRE_DEBUG.
784    
785        (b) In pcre_compile(), rename the member of the "branch_chain" structure
786            called "current" as "current_branch", to prevent a collision with the
787            Linux macro when compiled as a kernel module.
788    
789        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
790            prevent a collision with the Linux macro when compiled as a kernel
791            module.
792    
793    8.  In pcre_compile() there are some checks for integer overflows that used to
794        cast potentially large values to (double). This has been changed to that
795        when building, a check for int64_t is made, and if it is found, it is used
796        instead, thus avoiding the use of floating point arithmetic. (There is no
797        other use of FP in PCRE.) If int64_t is not found, the fallback is to
798        double.
799    
800    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
801        2005 (difference between two addresses compared to an unsigned value).
802    
803    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
804        custom one, because of the following reported problem in Windows:
805    
806          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
807              under Win32.
808          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
809              therefore missing the function definition.
810          - The compiler thus generates a "C" signature for the test function.
811          - The linker fails to find the "C" function.
812          - PCRE fails to configure if asked to do so against libbz2.
813    
814    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
815        messages were output:
816    
817          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
818          rerunning libtoolize, to keep the correct libtool macros in-tree.
819          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
820    
821        I have done both of these things.
822    
823    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
824        most of the time, it *can* run out if it is given a pattern that contains a
825        runaway infinite recursion. I updated the discussion in the pcrestack man
826        page.
827    
828    13. Now that we have gone to the x.xx style of version numbers, the minor
829        version may start with zero. Using 08 or 09 is a bad idea because users
830        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
831        interpreted as invalid octal numbers. I've updated the previous comment in
832        configure.ac, and also added a check that gives an error if 08 or 09 are
833        used.
834    
835    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
836        causing partial matching to fail when the end of the subject matched \W
837        in a UTF-8 pattern where \W was quantified with a minimum of 3.
838    
839    15. There were some discrepancies between the declarations in pcre_internal.h
840        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
841        their definitions. The declarations used "const uschar *" and the
842        definitions used USPTR. Even though USPTR is normally defined as "const
843        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
844        reported that: "This difference in casting confuses some C++ compilers, for
845        example, SunCC recognizes above declarations as different functions and
846        generates broken code for hbpcre." I have changed the declarations to use
847        USPTR.
848    
849    16. GNU libtool is named differently on some systems. The autogen.sh script now
850        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
851        (FreeBSD).
852    
853    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
854        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
855        comment: "Figure out how to create a longlong from a string: strtoll and
856        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
857        instance, but it only takes 2 args instead of 3!"
858    
859    18. A subtle bug concerned with back references has been fixed by a change of
860        specification, with a corresponding code fix. A pattern such as
861        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
862        refers, was giving matches when it shouldn't. For example, xa=xaaa would
863        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
864        same bug. Such groups have to be quantified to be useful, or contained
865        inside another quantified group. (If there's no repetition, the reference
866        can never match.) The problem arises because, having left the group and
867        moved on to the rest of the pattern, a later failure that backtracks into
868        the group uses the captured value from the final iteration of the group
869        rather than the correct earlier one. I have fixed this in PCRE by forcing
870        any group that contains a reference to itself to be an atomic group; that
871        is, there cannot be any backtracking into it once it has completed. This is
872        similar to recursive and subroutine calls.
873    
874    
875    Version 8.00 19-Oct-09
876    ----------------------
877    
878    1.  The table for translating pcre_compile() error codes into POSIX error codes
879        was out-of-date, and there was no check on the pcre_compile() error code
880        being within the table. This could lead to an OK return being given in
881        error.
882    
883    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
884        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
885        in a Windows environment.
886    
887    3.  The pcregrep --count option prints the count for each file even when it is
888        zero, as does GNU grep. However, pcregrep was also printing all files when
889        --files-with-matches was added. Now, when both options are given, it prints
890        counts only for those files that have at least one match. (GNU grep just
891        prints the file name in this circumstance, but including the count seems
892        more useful - otherwise, why use --count?) Also ensured that the
893        combination -clh just lists non-zero counts, with no names.
894    
895    4.  The long form of the pcregrep -F option was incorrectly implemented as
896        --fixed_strings instead of --fixed-strings. This is an incompatible change,
897        but it seems right to fix it, and I didn't think it was worth preserving
898        the old behaviour.
899    
900    5.  The command line items --regex=pattern and --regexp=pattern were not
901        recognized by pcregrep, which required --regex pattern or --regexp pattern
902        (with a space rather than an '='). The man page documented the '=' forms,
903        which are compatible with GNU grep; these now work.
904    
905    6.  No libpcreposix.pc file was created for pkg-config; there was just
906        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
907    
908    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
909        when UCP support is not needed, by modifying the Python script that
910        generates it from Unicode data files. This should not matter if the module
911        is correctly used as a library, but I received one complaint about 50K of
912        unwanted data. My guess is that the person linked everything into his
913        program rather than using a library. Anyway, it does no harm.
914    
915    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
916        was a minimum greater than 1 for a wide character in a possessive
917        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
918        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
919        character. Chaos in the form of incorrect output or a compiling loop could
920        result.
921    
922    9.  The restrictions on what a pattern can contain when partial matching is
923        requested for pcre_exec() have been removed. All patterns can now be
924        partially matched by this function. In addition, if there are at least two
925        slots in the offset vector, the offset of the earliest inspected character
926        for the match and the offset of the end of the subject are set in them when
927        PCRE_ERROR_PARTIAL is returned.
928    
929    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
930        synonymous with PCRE_PARTIAL, for backwards compatibility, and
931        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
932        and may be more useful for multi-segment matching.
933    
934    11. Partial matching with pcre_exec() is now more intuitive. A partial match
935        used to be given if ever the end of the subject was reached; now it is
936        given only if matching could not proceed because another character was
937        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
938        string "Z", which now yields "no match" instead of "partial match". In the
939        case of pcre_dfa_exec(), "no match" is given if every matching path for the
940        final character ended with (*FAIL).
941    
942    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
943        if the pattern had a "must contain" character that was already found in the
944        earlier partial match, unless partial matching was again requested. For
945        example, with the pattern /dog.(body)?/, the "must contain" character is
946        "g". If the first part-match was for the string "dog", restarting with
947        "sbody" failed. This bug has been fixed.
948    
949    13. The string returned by pcre_dfa_exec() after a partial match has been
950        changed so that it starts at the first inspected character rather than the
951        first character of the match. This makes a difference only if the pattern
952        starts with a lookbehind assertion or \b or \B (\K is not supported by
953        pcre_dfa_exec()). It's an incompatible change, but it makes the two
954        matching functions compatible, and I think it's the right thing to do.
955    
956    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
957        so that the demonstration program is easily available in environments where
958        PCRE has not been installed from source.
959    
960    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
961        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
962        library.
963    
964    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
965        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
966        is not the first non-POSIX option to be added. Clearly some people find
967        these options useful.
968    
969    17. If a caller to the POSIX matching function regexec() passes a non-zero
970        value for nmatch with a NULL value for pmatch, the value of
971        nmatch is forced to zero.
972    
973    18. RunGrepTest did not have a test for the availability of the -u option of
974        the diff command, as RunTest does. It now checks in the same way as
975        RunTest, and also checks for the -b option.
976    
977    19. If an odd number of negated classes containing just a single character
978        interposed, within parentheses, between a forward reference to a named
979        subpattern and the definition of the subpattern, compilation crashed with
980        an internal error, complaining that it could not find the referenced
981        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
982        [The bug was that it was starting one character too far in when skipping
983        over the character class, thus treating the ] as data rather than
984        terminating the class. This meant it could skip too much.]
985    
986    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
987        /g option in pcretest when the pattern contains \K, which makes it possible
988        to have an empty string match not at the start, even when the pattern is
989        anchored. Updated pcretest and pcredemo to use this option.
990    
991    21. If the maximum number of capturing subpatterns in a recursion was greater
992        than the maximum at the outer level, the higher number was returned, but
993        with unset values at the outer level. The correct (outer level) value is
994        now given.
995    
996    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
997        PCRE did not set those parentheses (unlike Perl). I have now found a way to
998        make it do so. The string so far is captured, making this feature
999        compatible with Perl.
1000    
1001    23. The tests have been re-organized, adding tests 11 and 12, to make it
1002        possible to check the Perl 5.10 features against Perl 5.10.
1003    
1004    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1005        pattern matches a fixed length string. PCRE did not allow this; now it
1006        does. Neither allows recursion.
1007    
1008    25. I finally figured out how to implement a request to provide the minimum
1009        length of subject string that was needed in order to match a given pattern.
1010        (It was back references and recursion that I had previously got hung up
1011        on.) This code has now been added to pcre_study(); it finds a lower bound
1012        to the length of subject needed. It is not necessarily the greatest lower
1013        bound, but using it to avoid searching strings that are too short does give
1014        some useful speed-ups. The value is available to calling programs via
1015        pcre_fullinfo().
1016    
1017    26. While implementing 25, I discovered to my embarrassment that pcretest had
1018        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1019        study optimizations had never been tested with that matching function.
1020        Oops. What is worse, even when it was passed study data, there was a bug in
1021        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1022        were also very few tests of studied patterns with pcre_dfa_exec().
1023    
1024    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1025        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1026        on the other side of the coin, they are no longer allowed to have different
1027        names, because these cannot be distinguished in PCRE, and this has caused
1028        confusion. (This is a difference from Perl.)
1029    
1030    28. When duplicate subpattern names are present (necessarily with different
1031        numbers, as required by 27 above), and a test is made by name in a
1032        conditional pattern, either for a subpattern having been matched, or for
1033        recursion in such a pattern, all the associated numbered subpatterns are
1034        tested, and the overall condition is true if the condition is true for any
1035        one of them. This is the way Perl works, and is also more like the way
1036        testing by number works.
1037    
1038    
1039    Version 7.9 11-Apr-09
1040    ---------------------
1041    
1042    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1043        (pcretest), all targets were linked against these libraries. This included
1044        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1045        libraries. This caused unwanted dependencies to be created. This problem
1046        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1047        pcretest is linked with readline.
1048    
1049    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1050        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1051        moved outside it again, because FALSE and TRUE are already defined in AIX,
1052        but BOOL is not.
1053    
1054    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1055        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1056    
1057    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1058        hyphens) following file names and line numbers when outputting matching
1059        lines. This is not true; no spaces are inserted. I have also clarified the
1060        wording for the --colour (or --color) option.
1061    
1062    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1063        was not coloured; this is different to GNU grep, so I have changed it to be
1064        the same.
1065    
1066    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1067        each matching line was coloured. Now it goes on to look for further matches
1068        of any of the test patterns, which is the same behaviour as GNU grep.
1069    
1070    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1071        doesn't make sense to accept an empty string match in pcregrep, so I have
1072        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1073        seems to be how GNU grep behaves.
1074    
1075    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1076        start or after a newline", because the conditional assertion was not being
1077        correctly handled. The rule now is that both the assertion and what follows
1078        in the first alternative must satisfy the test.
1079    
1080    9.  If auto-callout was enabled in a pattern with a conditional group whose
1081        condition was an assertion, PCRE could crash during matching, both with
1082        pcre_exec() and pcre_dfa_exec().
1083    
1084    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1085        used for matching.
1086    
1087    11. Unicode property support in character classes was not working for
1088        characters (bytes) greater than 127 when not in UTF-8 mode.
1089    
1090    12. Added the -M command line option to pcretest.
1091    
1092    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1093    
1094    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1095    
1096    16. Added comments and documentation about mis-use of no_arg in the C++
1097        wrapper.
1098    
1099    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1100        from Martin Jerabek that uses macro names for all relevant character and
1101        string constants.
1102    
1103    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1104        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1105        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1106        these, but not everybody uses configure.
1107    
1108    19. A conditional group that had only one branch was not being correctly
1109        recognized as an item that could match an empty string. This meant that an
1110        enclosing group might also not be so recognized, causing infinite looping
1111        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1112        with the subject "ab", where knowledge that the repeated group can match
1113        nothing is needed in order to break the loop.
1114    
1115    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1116        exec(), but without supplying a callout function, matching went wrong.
1117    
1118    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1119        leak if the size of the offset vector was greater than 30. When the vector
1120        is smaller, the saved offsets during recursion go onto a local stack
1121        vector, but for larger vectors malloc() is used. It was failing to free
1122        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1123        error, in fact).
1124    
1125    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1126        heapframe that is used only when UTF-8 support is enabled. This caused no
1127        problem, but was untidy.
1128    
1129    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1130        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1131        included within another project.
1132    
1133    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1134        slightly modified by me:
1135    
1136          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1137              not building pcregrep.
1138    
1139          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1140              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1141    
1142    25. Forward references, both numeric and by name, in patterns that made use of
1143        duplicate group numbers, could behave incorrectly or give incorrect errors,
1144        because when scanning forward to find the reference group, PCRE was not
1145        taking into account the duplicate group numbers. A pattern such as
1146        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1147    
1148    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1149        the feature of a custom pointer more persuasive (as requested by a user).
1150    
1151    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1152        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1153        pre-defined.
1154    
1155    28. Added support for (*UTF8) at the start of a pattern.
1156    
1157    29. Arrange for flags added by the "release type" setting in CMake to be shown
1158        in the configuration summary.
1159    
1160    
1161    Version 7.8 05-Sep-08
1162    ---------------------
1163    
1164    1.  Replaced UCP searching code with optimized version as implemented for Ad
1165        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1166        stage table and inline lookup instead of a function, giving speed ups of 2
1167        to 5 times on some simple patterns that I tested. Permission was given to
1168        distribute the MultiStage2.py script that generates the tables (it's not in
1169        the tarball, but is in the Subversion repository).
1170    
1171    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1172        scripts.
1173    
1174    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1175        a group with a zero qualifier. The result of the study could be incorrect,
1176        or the function might crash, depending on the pattern.
1177    
1178    4.  Caseless matching was not working for non-ASCII characters in back
1179        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1180        It now works when Unicode Property Support is available.
1181    
1182    5.  In pcretest, an escape such as \x{de} in the data was always generating
1183        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1184        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1185        truncation.
1186    
1187    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1188    
1189    7.  Added two (int) casts to pcregrep when printing the difference of two
1190        pointers, in case they are 64-bit values.
1191    
1192    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1193        test 2 if it fails.
1194    
1195    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1196        and a #define of that name to empty if it is not externally set. This is to
1197        allow users of MSVC to set it if necessary.
1198    
1199    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1200        the convenience functions in the pcre_get.c source file.
1201    
1202    11. An option change at the start of a pattern that had top-level alternatives
1203        could cause overwriting and/or a crash. This command provoked a crash in
1204        some environments:
1205    
1206          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1207    
1208        This potential security problem was recorded as CVE-2008-2371.
1209    
1210    12. For a pattern where the match had to start at the beginning or immediately
1211        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1212        pcre_dfa_exec() could read past the end of the passed subject if there was
1213        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1214        pcretest so that it places the subject at the end of its malloc-ed buffer.
1215    
1216    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1217        exec() might read past the end of the data buffer in UTF-8 mode.
1218    
1219    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1220        the data contained the byte 0x85 as part of a UTF-8 character within its
1221        first line. This applied both to normal and DFA matching.
1222    
1223    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1224        /^[^d]*?$/8 failed to match "abc".
1225    
1226    16. Added a missing copyright notice to pcrecpp_internal.h.
1227    
1228    17. Make it more clear in the documentation that values returned from
1229        pcre_exec() in ovector are byte offsets, not character counts.
1230    
1231    18. Tidied a few places to stop certain compilers from issuing warnings.
1232    
1233    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1234        supplied by Stefan Weber. I made a further small update for 7.8 because
1235        there is a change of source arrangements: the pcre_searchfuncs.c module is
1236        replaced by pcre_ucd.c.
1237    
1238    
1239    Version 7.7 07-May-08
1240    ---------------------
1241    
1242    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1243        a string to a long long, pretend we don't even have a long long." This is
1244        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1245    
1246    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1247        pre-7.6 versions, which defined a global no_arg variable instead of putting
1248        it in the RE class. (See also #8 below.)
1249    
1250    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1251        Lopes.
1252    
1253    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1254    
1255        (1) The include/exclude patterns were being applied to the whole pathnames
1256            of files, instead of just to the final components.
1257    
1258        (2) If there was more than one level of directory, the subdirectories were
1259            skipped unless they satisfied the include/exclude conditions. This is
1260            inconsistent with GNU grep (and could even be seen as contrary to the
1261            pcregrep specification - which I improved to make it absolutely clear).
1262            The action now is always to scan all levels of directory, and just
1263            apply the include/exclude patterns to regular files.
1264    
1265    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1266        --exclude_dir in the tests to avoid scanning .svn directories.
1267    
1268    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1269        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1270        doesn't support NULs in patterns.
1271    
1272    7.  Added some missing "const"s to declarations of static tables in
1273        pcre_compile.c and pcre_dfa_exec.c.
1274    
1275    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1276        caused by fix #2  above. (Subsequently also a second patch to fix the
1277        first patch. And a third patch - this was a messy problem.)
1278    
1279    9.  Applied Craig's patch to remove the use of push_back().
1280    
1281    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1282        matching function regexec().
1283    
1284    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1285        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1286        references. PCRE supports relative numbers with this syntax (I don't think
1287        Oniguruma does).
1288    
1289    12. Previously, a group with a zero repeat such as (...){0} was completely
1290        omitted from the compiled regex. However, this means that if the group
1291        was called as a subroutine from elsewhere in the pattern, things went wrong
1292        (an internal error was given). Such groups are now left in the compiled
1293        pattern, with a new opcode that causes them to be skipped at execution
1294        time.
1295    
1296    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1297        to the way PCRE behaves:
1298    
1299        (a) A lone ] character is dis-allowed (Perl treats it as data).
1300    
1301        (b) A back reference to an unmatched subpattern matches an empty string
1302            (Perl fails the current match path).
1303    
1304        (c) A data ] in a character class must be notated as \] because if the
1305            first data character in a class is ], it defines an empty class. (In
1306            Perl it is not possible to have an empty class.) The empty class []
1307            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1308            The negative empty class [^] matches any one character, independently
1309            of the DOTALL setting.
1310    
1311    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1312        non-existent subpattern following a character class starting with ']' and
1313        containing () gave an internal compiling error instead of "reference to
1314        non-existent subpattern". Fortunately, when the pattern did exist, the
1315        compiled code was correct. (When scanning forwards to check for the
1316        existencd of the subpattern, it was treating the data ']' as terminating
1317        the class, so got the count wrong. When actually compiling, the reference
1318        was subsequently set up correctly.)
1319    
1320    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1321        it was being rejected as not supported by pcre_dfa_exec(), even though
1322        other assertions are supported. I have made pcre_dfa_exec() support
1323        (*FAIL).
1324    
1325    16. The implementation of 13c above involved the invention of a new opcode,
1326        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1327        cannot be changed at match time, I realized I could make a small
1328        improvement to matching performance by compiling OP_ALLANY instead of
1329        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1330        on the OP_ANY path.
1331    
1332    17. Compiling pcretest on Windows with readline support failed without the
1333        following two fixes: (1) Make the unistd.h include conditional on
1334        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1335    
1336    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1337        ncurses library to be included for pcretest when ReadLine support is
1338        requested, but also to allow for it to be overridden. This patch came from
1339        Daniel Bergström.
1340    
1341    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1342        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1343        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1344        spotting this.
1345    
1346    
1347    Version 7.6 28-Jan-08
1348    ---------------------
1349    
1350    1.  A character class containing a very large number of characters with
1351        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1352        overflow.
1353    
1354    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1355        HAVE_LONG_LONG is not defined.
1356    
1357    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1358        bring them up to date and include new features. This patch includes:
1359    
1360        - Fixed PH's badly added libz and libbz2 support.
1361        - Fixed a problem with static linking.
1362        - Added pcredemo. [But later removed - see 7 below.]
1363        - Fixed dftables problem and added an option.
1364        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1365            HAVE_LONG_LONG.
1366        - Added readline support for pcretest.
1367        - Added an listing of the option settings after cmake has run.
1368    
1369    4.  A user submitted a patch to Makefile that makes it easy to create
1370        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1371        Makefile.am that cause it to include this special target, without
1372        affecting anything else. Note that the same mingw target plus all
1373        the other distribution libraries and programs are now supported
1374        when configuring with CMake (see 6 below) instead of with
1375        Configure/Make.
1376    
1377    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1378        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1379        exported in the Windows port". It has not yet been confirmed that the patch
1380        solves the problem, but it does no harm.
1381    
1382    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1383        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1384        with CMake, and also correct the comment about stack recursion.
1385    
1386    7.  Remove the automatic building of pcredemo from the ./configure system and
1387        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1388        of a program that users should build themselves after PCRE is installed, so
1389        building it automatically is not really right. What is more, it gave
1390        trouble in some build environments.
1391    
1392    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1393    
1394    
1395    Version 7.5 10-Jan-08
1396    ---------------------
1397    
1398    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1399        values in parens when parsing an RE using the C++ wrapper."
1400    
1401    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1402        Characters greater than 255 were excluded from the class instead of being
1403        included.
1404    
1405    3.  The same bug as (2) above applied to negated POSIX classes such as
1406        [:^space:].
1407    
1408    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1409        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1410        I have changed it.
1411    
1412    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1413        first named subpattern) and a construct such as (?&a) would reference the
1414        first named subpattern whose name started with "a" (in other words, the
1415        length check was missing). Both these problems are fixed. "Subpattern name
1416        expected" is now given for (?&) (a zero-length name), and this patch also
1417        makes it give the same error for \k'' (previously it complained that that
1418        was a reference to a non-existent subpattern).
1419    
1420    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1421        this is right because (?- can be followed by option settings as well as by
1422        digits. I have, however, made the messages clearer.
1423    
1424    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1425        than the number used in the conditional) now cause a compile-time error.
1426        This is actually not compatible with Perl, which accepts such patterns, but
1427        treats the conditional as always being FALSE (as PCRE used to), but it
1428        seems to me that giving a diagnostic is better.
1429    
1430    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1431        and messages.
1432    
1433    9.  Fix two occurrences of "backslash" in comments that should have been
1434        "backspace".
1435    
1436    10. Remove two redundant lines of code that can never be obeyed (their function
1437        was moved elsewhere).
1438    
1439    11. The program that makes PCRE's Unicode character property table had a bug
1440        which caused it to generate incorrect table entries for sequences of
1441        characters that have the same character type, but are in different scripts.
1442        It amalgamated them into a single range, with the script of the first of
1443        them. In other words, some characters were in the wrong script. There were
1444        thirteen such cases, affecting characters in the following ranges:
1445    
1446          U+002b0 - U+002c1
1447          U+0060c - U+0060d
1448          U+0061e - U+00612
1449          U+0064b - U+0065e
1450          U+0074d - U+0076d
1451          U+01800 - U+01805
1452          U+01d00 - U+01d77
1453          U+01d9b - U+01dbf
1454          U+0200b - U+0200f
1455          U+030fc - U+030fe
1456          U+03260 - U+0327f
1457          U+0fb46 - U+0fbb1
1458          U+10450 - U+1049d
1459    
1460    12. The -o option (show only the matching part of a line) for pcregrep was not
1461        compatible with GNU grep in that, if there was more than one match in a
1462        line, it showed only the first of them. It now behaves in the same way as
1463        GNU grep.
1464    
1465    13. If the -o and -v options were combined for pcregrep, it printed a blank
1466        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1467        does the same. The return code can be used to tell if there were any
1468        non-matching lines.
1469    
1470    14. Added --file-offsets and --line-offsets to pcregrep.
1471    
1472    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1473        infinitely looping recursion. The bug was that positive lookaheads were not
1474        being skipped when checking for a possible empty match (negative lookaheads
1475        and both kinds of lookbehind were skipped).
1476    
1477    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1478        inclusion of <windows.h> to before rather than after the definition of
1479        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1480    
1481    17. Specifying a possessive quantifier with a specific limit for a Unicode
1482        character property caused pcre_compile() to compile bad code, which led at
1483        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1484        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1485        caused the error; without that there was no problem.
1486    
1487    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1488    
1489    19. Added --enable-pcretest-libreadline.
1490    
1491    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1492        RE::GlobalReplace(). As a result, the number of replacements returned was
1493        double what it should be. I removed one of the increments, but Craig sent a
1494        later patch that removed the other one (the right fix) and added unit tests
1495        that check the return values (which was not done before).
1496    
1497    21. Several CMake things:
1498    
1499        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1500            the names libpcre and libpcreposix, not just pcre and pcreposix.
1501    
1502        (2) The above change means that pcretest and pcregrep are now correctly
1503            linked with the newly-built libraries, not previously installed ones.
1504    
1505        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1506    
1507    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1508        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1509        UTF-8 newline character). The key issue is that the pattern starts .*;
1510        this means that the match must be either at the beginning, or after a
1511        newline. The bug was in the code for advancing after a failed match and
1512        checking that the new position followed a newline. It was not taking
1513        account of UTF-8 characters correctly.
1514    
1515    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1516        character classes. PCRE was not treating the sequence [:...:] as a
1517        character class unless the ... were all letters. Perl, however, seems to
1518        allow any characters between [: and :], though of course it rejects as
1519        unknown any "names" that contain non-letters, because all the known class
1520        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1521        for example, whereas PCRE did not - it did not recognize a POSIX character
1522        class. This seemed a bit dangerous, so the code has been changed to be
1523        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1524        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1525        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1526        Perl does, and where it didn't before.
1527    
1528    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1529        Windows environments %n is disabled by default.
1530    
1531    
1532    Version 7.4 21-Sep-07
1533    ---------------------
1534    
1535    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1536        means that a class such as [\s] counted as "explicit reference to CR or
1537        LF". That isn't really right - the whole point of the change was to try to
1538        help when there was an actual mention of one of the two characters. So now
1539        the change happens only if \r or \n (or a literal CR or LF) character is
1540        encountered.
1541    
1542    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1543        of both had grown to the point where there were only 3 bits left.
1544        Fortunately, there was spare space in the data structure, and so I have
1545        moved the internal flags into a new 16-bit field to free up more option
1546        bits.
1547    
1548    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1549        but did not set the internal JCHANGED flag - either of these is enough to
1550        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1551        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1552        start sets both bits.
1553    
1554    4.  Added options (at build time, compile time, exec time) to change \R from
1555        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1556    
1557    5.  doc/pcresyntax.html was missing from the distribution.
1558    
1559    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1560        compatibility, even though it is no longer used.
1561    
1562    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1563        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1564        windows.h file is present (where different names are used). [This was
1565        reversed later after testing - see 16 below.]
1566    
1567    8.  Changed all #include <config.h> to #include "config.h". There were also
1568        some further <pcre.h> cases that I changed to "pcre.h".
1569    
1570    9.  When pcregrep was used with the --colour option, it missed the line ending
1571        sequence off the lines that it output.
1572    
1573    10. It was pointed out to me that arrays of string pointers cause lots of
1574        relocations when a shared library is dynamically loaded. A technique of
1575        using a single long string with a table of offsets can drastically reduce
1576        these. I have refactored PCRE in four places to do this. The result is
1577        dramatic:
1578    
1579          Originally:                          290
1580          After changing UCP table:            187
1581          After changing error message table:   43
1582          After changing table of "verbs"       36
1583          After changing table of Posix names   22
1584    
1585        Thanks to the folks working on Gregex for glib for this insight.
1586    
1587    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1588        unicode-properties was also set.
1589    
1590    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1591    
1592    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1593        checked only for CRLF.
1594    
1595    14. Added casts to pcretest.c to avoid compiler warnings.
1596    
1597    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1598    
1599    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1600        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1601        entirely. This removes changes made in 7 above.
1602    
1603    17. The CMake files have been updated, and there is now more information about
1604        building with CMake in the NON-UNIX-USE document.
1605    
1606    
1607    Version 7.3 28-Aug-07
1608    ---------------------
1609    
1610     1. In the rejigging of the build system that eventually resulted in 7.1, the
1611        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1612        brackets there is not right, since it causes compilers to look for an
1613        installed pcre.h, not the version that is in the source that is being
1614        compiled (which of course may be different). I have changed it back to:
1615    
1616          #include "pcre.h"
1617    
1618        I have a vague recollection that the change was concerned with compiling in
1619        different directories, but in the new build system, that is taken care of
1620        by the VPATH setting the Makefile.
1621    
1622     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1623        when the subject happened to end in the byte 0x85 (e.g. if the last
1624        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1625        characters but of course it shouldn't be taken as a newline when it is part
1626        of another character. The bug was that, for an unlimited repeat of . in
1627        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1628        characters when looking for a newline.
1629    
1630     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1631    
1632     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1633        in debug output.
1634    
1635     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1636        long printing in the pcrecpp unittest when running under MinGW.
1637    
1638     6. ESC_K was left out of the EBCDIC table.
1639    
1640     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1641        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1642        limit also applies to "virtual nesting" when a pattern is recursive, and in
1643        this case 1000 isn't so big. I have been able to remove this limit at the
1644        expense of backing off one optimization in certain circumstances. Normally,
1645        when pcre_exec() would call its internal match() function recursively and
1646        immediately return the result unconditionally, it uses a "tail recursion"
1647        feature to save stack. However, when a subpattern that can match an empty
1648        string has an unlimited repetition quantifier, it no longer makes this
1649        optimization. That gives it a stack frame in which to save the data for
1650        checking that an empty string has been matched. Previously this was taken
1651        from the 1000-entry workspace that had been reserved. So now there is no
1652        explicit limit, but more stack is used.
1653    
1654     8. Applied Daniel's patches to solve problems with the import/export magic
1655        syntax that is required for Windows, and which was going wrong for the
1656        pcreposix and pcrecpp parts of the library. These were overlooked when this
1657        problem was solved for the main library.
1658    
1659     9. There were some crude static tests to avoid integer overflow when computing
1660        the size of patterns that contain repeated groups with explicit upper
1661        limits. As the maximum quantifier is 65535, the maximum group length was
1662        set at 30,000 so that the product of these two numbers did not overflow a
1663        32-bit integer. However, it turns out that people want to use groups that
1664        are longer than 30,000 bytes (though not repeat them that many times).
1665        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1666        made it possible to implement the integer overflow checks in a much more
1667        dynamic way, which I have now done. The artificial limitation on group
1668        length has been removed - we now have only the limit on the total length of
1669        the compiled pattern, which depends on the LINK_SIZE setting.
1670    
1671    10. Fixed a bug in the documentation for get/copy named substring when
1672        duplicate names are permitted. If none of the named substrings are set, the
1673        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1674        empty string.
1675    
1676    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1677        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1678        because the ] is interpreted as the first data character and the
1679        terminating ] is not found. PCRE has been made compatible with Perl in this
1680        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1681        cause memory overwriting.
1682    
1683    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1684        string has been matched (to stop an infinite loop). It was not recognizing
1685        a conditional subpattern that could match an empty string if that
1686        subpattern was within another subpattern. For example, it looped when
1687        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1688        condition was not nested. This bug has been fixed.
1689    
1690    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1691        past the start of the subject in the presence of bytes with the top bit
1692        set, for example "\x8aBCD".
1693    
1694    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1695        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1696    
1697    14. Optimized (?!) to (*FAIL).
1698    
1699    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1700        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1701        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1702        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1703        does: it's just the validity check that is more restrictive.
1704    
1705    16. Inserted checks for integer overflows during escape sequence (backslash)
1706        processing, and also fixed erroneous offset values for syntax errors during
1707        backslash processing.
1708    
1709    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1710        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1711    
1712    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1713        caused an overrun.
1714    
1715    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1716        something other than just ASCII characters) inside a group that had an
1717        unlimited repeat caused a loop at compile time (while checking to see
1718        whether the group could match an empty string).
1719    
1720    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1721        [\P{Any}] did so. (Error in the code for printing property names.)
1722    
1723    21. An orphan \E inside a character class could cause a crash.
1724    
1725    22. A repeated capturing bracket such as (A)? could cause a wild memory
1726        reference during compilation.
1727    
1728    23. There are several functions in pcre_compile() that scan along a compiled
1729        expression for various reasons (e.g. to see if it's fixed length for look
1730        behind). There were bugs in these functions when a repeated \p or \P was
1731        present in the pattern. These operators have additional parameters compared
1732        with \d, etc, and these were not being taken into account when moving along
1733        the compiled data. Specifically:
1734    
1735        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1736            length.
1737    
1738        (b) An item such as \pL+ within a repeated group could cause crashes or
1739            loops.
1740    
1741        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1742            "reference to non-existent subpattern" error.
1743    
1744        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1745    
1746    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1747        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1748    
1749    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1750    
1751    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1752        character were causing crashes (broken optimization).
1753    
1754    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1755        \p or \P) caused a compile-time loop.
1756    
1757    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1758        break. For example, the unstudied pattern [\r\n]A does not match the string
1759        "\r\nA" because change 7.0/46 below moves the current point on by two
1760        characters after failing to match at the start. However, the pattern \nA
1761        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1762        the same is true. There doesn't seem any very clean way out of this, but
1763        what I have chosen to do makes the common cases work: PCRE now takes note
1764        of whether there can be an explicit match for \r or \n anywhere in the
1765        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1766        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1767        pattern has explicit CR or LF references.
1768    
1769    29. Added (*CR) etc for changing newline setting at start of pattern.
1770    
1771    
1772    Version 7.2 19-Jun-07
1773    ---------------------
1774    
1775     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1776        which is apparently normally available under Windows.
1777    
1778     2. Re-jig the pcregrep tests with different newline settings in an attempt
1779        to make them independent of the local environment's newline setting.
1780    
1781     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1782    
1783     4. Some of the "internals" tests were previously cut out when the link size
1784        was not 2, because the output contained actual offsets. The recent new
1785        "Z" feature of pcretest means that these can be cut out, making the tests
1786        usable with all link sizes.
1787    
1788     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1789        stack recursion. This gives a massive performance boost under BSD, but just
1790        a small improvement under Linux. However, it saves one field in the frame
1791        in all cases.
1792    
1793     6. Added more features from the forthcoming Perl 5.10:
1794    
1795        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1796            recursion call. It refers to the nth most recently opened parentheses.
1797    
1798        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1799            to be opened parentheses.
1800    
1801        (c) Conditions that refer to capturing parentheses can be specified
1802            relatively, for example, (?(-2)... or (?(+3)...
1803    
1804        (d) \K resets the start of the current match so that everything before
1805            is not part of it.
1806    
1807        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1808    
1809        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1810            reference syntax.
1811    
1812        (g) (?| introduces a group in which the numbering of parentheses in each
1813            alternative starts with the same number.
1814    
1815        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1816    
1817     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1818        PCRE_INFO_JCHANGED.
1819    
1820     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1821        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1822        for detecting groups that can match an empty string.
1823    
1824     9. A pattern with a very large number of alternatives (more than several
1825        hundred) was running out of internal workspace during the pre-compile
1826        phase, where pcre_compile() figures out how much memory will be needed. A
1827        bit of new cunning has reduced the workspace needed for groups with
1828        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1829        workspace instead of running out of the 4096 that are available.
1830    
1831    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1832    
1833    11. Applied patch from Google to remove an optimization that didn't quite work.
1834        The report of the bug said:
1835    
1836          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1837          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1838          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1839    
1840    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1841        it matched the wrong number of bytes.
1842    
1843    
1844    Version 7.1 24-Apr-07
1845    ---------------------
1846    
1847     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1848        that is more "standard", making use of automake and other Autotools. There
1849        is some re-arrangement of the files and adjustment of comments consequent
1850        on this.
1851    
1852     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1853        for recursive directory scanning broke on some systems because the files
1854        are not scanned in any specific order and on different systems the order
1855        was different. A call to "sort" has been inserted into RunGrepTest for the
1856        approprate test as a short-term fix. In the longer term there may be an
1857        alternative.
1858    
1859     3. I had an email from Eric Raymond about problems translating some of PCRE's
1860        man pages to HTML (despite the fact that I distribute HTML pages, some
1861        people do their own conversions for various reasons). The problems
1862        concerned the use of low-level troff macros .br and .in. I have therefore
1863        removed all such uses from the man pages (some were redundant, some could
1864        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1865        HTML has been updated to handle .nf/.fi and to complain if it encounters
1866        .br or .in.
1867    
1868     4. Updated comments in configure.ac that get placed in config.h.in and also
1869        arranged for config.h to be included in the distribution, with the name
1870        config.h.generic, for the benefit of those who have to compile without
1871        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1872    
1873     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1874        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1875        makevp.bat for the current PCRE, using the additional files
1876        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1877    
1878     6. A Windows user reported a minor discrepancy with test 2, which turned out
1879        to be caused by a trailing space on an input line that had got lost in his
1880        copy. The trailing space was an accident, so I've just removed it.
1881    
1882     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1883        that is needed.
1884    
1885     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1886        as "const" (a) because they are and (b) because it helps the PHP
1887        maintainers who have recently made a script to detect big data structures
1888        in the php code that should be moved to the .rodata section. I remembered
1889        to update Builducptable as well, so it won't revert if ucptable.h is ever
1890        re-created.
1891    
1892     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1893        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1894        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1895        support is not required. This saves 1.5-2K of code, which is important in
1896        some applications.
1897    
1898        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1899        so as not to refer to the tables, even though these functions will never be
1900        called when UTF-8 support is disabled. Otherwise there are problems with a
1901        shared library.
1902    
1903    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1904    
1905        (a) It was defining its arguments as char * instead of void *.
1906    
1907        (b) It was assuming that all moves were upwards in memory; this was true
1908            a long time ago when I wrote it, but is no longer the case.
1909    
1910        The emulated memove() is provided for those environments that have neither
1911        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1912        is clearly not the case, as these two bugs were recently reported.
1913    
1914    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1915        and Detrail to create the HTML documentation, the .txt form of the man
1916        pages, and it removes trailing spaces from listed files. It also creates
1917        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1918        case, it wraps all the #defines with #ifndefs. This script should be run
1919        before "make dist".
1920    
1921    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1922        with Unicode property support.
1923    
1924        (a) For a maximizing quantifier, if the two different cases of the
1925            character were of different lengths in their UTF-8 codings (there are
1926            some cases like this - I found 11), and the matching function had to
1927            back up over a mixture of the two cases, it incorrectly assumed they
1928            were both the same length.
1929    
1930        (b) When PCRE was configured to use the heap rather than the stack for
1931            recursion during matching, it was not correctly preserving the data for
1932            the other case of a UTF-8 character when checking ahead for a match
1933            while processing a minimizing repeat. If the check also involved
1934            matching a wide character, but failed, corruption could cause an
1935            erroneous result when trying to check for a repeat of the original
1936            character.
1937    
1938    13. Some tidying changes to the testing mechanism:
1939    
1940        (a) The RunTest script now detects the internal link size and whether there
1941            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1942            values substituted by "configure". (The RunGrepTest script already did
1943            this for UTF-8.) The configure.ac script no longer substitutes the
1944            relevant variables.
1945    
1946        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1947            with length and offset values. This means that the output is different
1948            for different internal link sizes. Test 2 is skipped for link sizes
1949            other than 2 because of this, bypassing the problem. Unfortunately,
1950            there was also a test in test 3 (the locale tests) that used /B and
1951            failed for link sizes other than 2. Rather than cut the whole test out,
1952            I have added a new /Z option to pcretest that replaces the length and
1953            offset values with spaces. This is now used to make test 3 independent
1954            of link size. (Test 2 will be tidied up later.)
1955    
1956    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1957        segmentation fault instead of returning the appropriate error message.
1958    
1959    15. In multiline mode when the newline sequence was set to "any", the pattern
1960        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1961        This doesn't seem right; it now treats the CRLF combination as the line
1962        ending, and so does not match in that case. It's only a pattern such as ^$
1963        that would hit this one: something like ^ABC$ would have failed after \r
1964        and then tried again after \r\n.
1965    
1966    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1967        in an attempt to make files that differ only in their line terminators
1968        compare equal. This works on Linux.
1969    
1970    17. Under certain error circumstances pcregrep might try to free random memory
1971        as it exited. This is now fixed, thanks to valgrind.
1972    
1973    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1974        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1975        was because its rules for how to advance for /g after matching an empty
1976        string at the end of a line did not allow for this case. They now check for
1977        it specially.
1978    
1979    20. pcretest is supposed to handle patterns and data of any length, by
1980        extending its buffers when necessary. It was getting this wrong when the
1981        buffer for a data line had to be extended.
1982    
1983    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1984        CRLF as a newline sequence.
1985    
1986    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1987        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1988        I have nevertheless tidied it up.
1989    
1990    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1991    
1992    24. Added a man page for pcre-config.
1993    
1994    
1995    Version 7.0 19-Dec-06
1996    ---------------------
1997    
1998     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1999        moving to gcc 4.1.1.
2000    
2001     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2002        sys/time.h, which is documented as needed for this function. It doesn't
2003        seem to matter on Linux, but it showed up on some releases of OS X.
2004    
2005     3. It seems that there are systems where bytes whose values are greater than
2006        127 match isprint() in the "C" locale. The "C" locale should be the
2007        default when a C program starts up. In most systems, only ASCII printing
2008        characters match isprint(). This difference caused the output from pcretest
2009        to vary, making some of the tests fail. I have changed pcretest so that:
2010    
2011        (a) When it is outputting text in the compiled version of a pattern, bytes
2012            other than 32-126 are always shown as hex escapes.
2013    
2014        (b) When it is outputting text that is a matched part of a subject string,
2015            it does the same, unless a different locale has been set for the match
2016            (using the /L modifier). In this case, it uses isprint() to decide.
2017    
2018     4. Fixed a major bug that caused incorrect computation of the amount of memory
2019        required for a compiled pattern when options that changed within the
2020        pattern affected the logic of the preliminary scan that determines the
2021        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2022        that the computed length was too small. The symptoms of this bug were
2023        either the PCRE error "internal error: code overflow" from pcre_compile(),
2024        or a glibc crash with a message such as "pcretest: free(): invalid next
2025        size (fast)". Examples of patterns that provoked this bug (shown in
2026        pcretest format) are:
2027    
2028          /(?-x: )/x
2029          /(?x)(?-x: \s*#\s*)/
2030          /((?i)[\x{c0}])/8
2031          /(?i:[\x{c0}])/8
2032    
2033        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2034        is now done differently.
2035    
2036     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2037        wrapper classes; (b) implement a new function in the C++ scanner that is
2038        more efficient than the old way of doing things because it avoids levels of
2039        recursion in the regex matching; (c) add a paragraph to the documentation
2040        for the FullMatch() function.
2041    
2042     6. The escape sequence \n was being treated as whatever was defined as
2043        "newline". Not only was this contrary to the documentation, which states
2044        that \n is character 10 (hex 0A), but it also went horribly wrong when
2045        "newline" was defined as CRLF. This has been fixed.
2046    
2047     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2048        was being set to -1 for the "end of line" case (supposedly a value that no
2049        character can have). Though this value is never used (the check for end of
2050        line is "zero bytes in current character"), it caused compiler complaints.
2051        I've changed it to 0xffffffff.
2052    
2053     8. In pcre_version.c, the version string was being built by a sequence of
2054        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2055        string (as it is for production releases) called a macro with an empty
2056        argument. The C standard says the result of this is undefined. The gcc
2057        compiler treats it as an empty string (which was what was wanted) but it is
2058        reported that Visual C gives an error. The source has been hacked around to
2059        avoid this problem.
2060    
2061     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2062        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2063        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2064        of them did).
2065    
2066    10. Originally, pcretest opened its input and output without "b"; then I was
2067        told that "b" was needed in some environments, so it was added for release
2068        5.0 to both the input and output. (It makes no difference on Unix-like
2069        systems.) Later I was told that it is wrong for the input on Windows. I've
2070        now abstracted the modes into two macros, to make it easier to fiddle with
2071        them, and removed "b" from the input mode under Windows.
2072    
2073    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2074    
2075    12. Added -help and --help to pcretest as an official way of being reminded
2076        of the options.
2077    
2078    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2079        and pcrecpp.cc because they annoy compilers at high warning levels.
2080    
2081    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2082    
2083    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2084        scripts are not C programs :-) and which was not noticed because it works
2085        on Linux.
2086    
2087    16. pcretest is supposed to handle any length of pattern and data line (as one
2088        line or as a continued sequence of lines) by extending its input buffer if
2089        necessary. This feature was broken for very long pattern lines, leading to
2090        a string of junk being passed to pcre_compile() if the pattern was longer
2091        than about 50K.
2092    
2093    17. I have done a major re-factoring of the way pcre_compile() computes the
2094        amount of memory needed for a compiled pattern. Previously, there was code
2095        that made a preliminary scan of the pattern in order to do this. That was
2096        OK when PCRE was new, but as the facilities have expanded, it has become
2097        harder and harder to keep it in step with the real compile phase, and there
2098        have been a number of bugs (see for example, 4 above). I have now found a
2099        cunning way of running the real compile function in a "fake" mode that
2100        enables it to compute how much memory it would need, while actually only
2101        ever using a few hundred bytes of working memory and without too many
2102        tests of the mode. This should make future maintenance and development
2103        easier. A side effect of this work is that the limit of 200 on the nesting
2104        depth of parentheses has been removed (though this was never a serious
2105        limitation, I suspect). However, there is a downside: pcre_compile() now
2106        runs more slowly than before (30% or more, depending on the pattern). I
2107        hope this isn't a big issue. There is no effect on runtime performance.
2108    
2109    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2110        newline (only possible for the last line of a file) and it was a
2111        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2112    
2113    19. Added additional timing features to pcretest. (1) The -tm option now times
2114        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2115        separate command line item, by a number that specifies the number of
2116        repeats to use when timing. The default is 50000; this gives better
2117        precision, but takes uncomfortably long for very large patterns.
2118    
2119    20. Extended pcre_study() to be more clever in cases where a branch of a
2120        subpattern has no definite first character. For example, (a*|b*)[cd] would
2121        previously give no result from pcre_study(). Now it recognizes that the
2122        first character must be a, b, c, or d.
2123    
2124    21. There was an incorrect error "recursive call could loop indefinitely" if
2125        a subpattern (or the entire pattern) that was being tested for matching an
2126        empty string contained only one non-empty item after a nested subpattern.
2127        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2128        incorrectly, because the \d was being skipped in the check.
2129    
2130    22. The pcretest program now has a new pattern option /B and a command line
2131        option -b, which is equivalent to adding /B to every pattern. This causes
2132        it to show the compiled bytecode, without the additional information that
2133        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2134        is the same as /B/I).
2135    
2136    23. A new optimization is now able automatically to treat some sequences such
2137        as a*b as a*+b. More specifically, if something simple (such as a character
2138        or a simple class like \d) has an unlimited quantifier, and is followed by
2139        something that cannot possibly match the quantified thing, the quantifier
2140        is automatically "possessified".
2141    
2142    24. A recursive reference to a subpattern whose number was greater than 39
2143        went wrong under certain circumstances in UTF-8 mode. This bug could also
2144        have affected the operation of pcre_study().
2145    
2146    25. Realized that a little bit of performance could be had by replacing
2147        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2148    
2149    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2150    
2151    27. Possessive quantifiers such as a++ were previously implemented by turning
2152        them into atomic groups such as ($>a+). Now they have their own opcodes,
2153        which improves performance. This includes the automatically created ones
2154        from 23 above.
2155    
2156    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2157        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2158        the first matched character to be a colon. This applied both to named and
2159        numbered groups.
2160    
2161    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2162    
2163    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2164        building PCRE on Windows easier, so I have included it in the distribution.
2165    
2166    31. There is now a check in pcretest against a ridiculously large number being
2167        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2168        loop, the loop is abandoned.
2169    
2170    32. Forward references to subpatterns in conditions such as (?(2)...) where
2171        subpattern 2 is defined later cause pcre_compile() to search forwards in
2172        the pattern for the relevant set of parentheses. This search went wrong
2173        when there were unescaped parentheses in a character class, parentheses
2174        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2175    
2176    33. "Subroutine" calls and backreferences were previously restricted to
2177        referencing subpatterns earlier in the regex. This restriction has now
2178        been removed.
2179    
2180    34. Added a number of extra features that are going to be in Perl 5.10. On the
2181        whole, these are just syntactic alternatives for features that PCRE had
2182        previously implemented using the Python syntax or my own invention. The
2183        other formats are all retained for compatibility.
2184    
2185        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2186            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2187            also .NET compatible.
2188    
2189        (b) A recursion or subroutine call to a named group can now be defined as
2190            (?&name) as well as (?P>name).
2191    
2192        (c) A backreference to a named group can now be defined as \k<name> or
2193            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2194            5.10, are also .NET compatible.
2195    
2196        (d) A conditional reference to a named group can now use the syntax
2197            (?(<name>) or (?('name') as well as (?(name).
2198    
2199        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2200            groups (named and numbered) that are never evaluated inline, but can be
2201            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2202            is always false. There may be only one alternative in such a group.
2203    
2204        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2205            as the simple (?(R). The condition is true only if the most recent
2206            recursion is that of the given number or name. It does not search out
2207            through the entire recursion stack.
2208    
2209        (g) The escape \gN or \g{N} has been added, where N is a positive or
2210            negative number, specifying an absolute or relative reference.
2211    
2212    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2213        some "unreachable code" warnings.
2214    
2215    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2216        things, this adds five new scripts.
2217    
2218    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2219        There were also incompatibilities regarding the handling of \Q..\E inside
2220        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2221        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2222    
2223    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2224        matches an empty string, and forcibly breaks the loop. There were bugs in
2225        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2226        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2227        separate and independent bugs (that affected different cases) have been
2228        fixed.
2229    
2230    39. Refactored the code to abolish the use of different opcodes for small
2231        capturing bracket numbers. This is a tidy that I avoided doing when I
2232        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2233        The new approach is not only tidier, it makes it possible to reduce the
2234        memory needed to fix the previous bug (38).
2235    
2236    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2237        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2238        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2239        mode.
2240    
2241    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2242        report.
2243    
2244    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2245        copy construction and assignment in the C++ wrapper.
2246    
2247    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2248        couple of bugs that could have given wrong results in the "--newline=crlf"
2249        case.
2250    
2251    44. Added a number of casts and did some reorganization of signed/unsigned int
2252        variables following suggestions from Dair Grant. Also renamed the variable
2253        "this" as "item" because it is a C++ keyword.
2254    
2255    45. Arranged for dftables to add
2256    
2257          #include "pcre_internal.h"
2258    
2259        to pcre_chartables.c because without it, gcc 4.x may remove the array
2260        definition from the final binary if PCRE is built into a static library and
2261        dead code stripping is activated.
2262    
2263    46. For an unanchored pattern, if a match attempt fails at the start of a
2264        newline sequence, and the newline setting is CRLF or ANY, and the next two
2265        characters are CRLF, advance by two characters instead of one.
2266    
2267    
2268    Version 6.7 04-Jul-06
2269    ---------------------
2270    
2271     1. In order to handle tests when input lines are enormously long, pcretest has
2272        been re-factored so that it automatically extends its buffers when
2273        necessary. The code is crude, but this _is_ just a test program. The
2274        default size has been increased from 32K to 50K.
2275    
2276     2. The code in pcre_study() was using the value of the re argument before
2277        testing it for NULL. (Of course, in any sensible call of the function, it
2278        won't be NULL.)
2279    
2280     3. The memmove() emulation function in pcre_internal.h, which is used on
2281        systems that lack both memmove() and bcopy() - that is, hardly ever -
2282        was missing a "static" storage class specifier.
2283    
2284     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2285        containing an extended class (one that cannot be represented by a bitmap
2286        because it contains high-valued characters or Unicode property items, e.g.
2287        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2288        pattern, but PCRE should not loop if you do not (it no longer does).
2289        [Detail: two cases were found: (a) a repeated subpattern containing an
2290        extended class; (b) a recursive reference to a subpattern that followed a
2291        previous extended class. It wasn't skipping over the extended class
2292        correctly when UTF-8 mode was not set.]
2293    
2294     5. A negated single-character class was not being recognized as fixed-length
2295        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2296        compile error "lookbehind assertion is not fixed length".
2297    
2298     6. The RunPerlTest auxiliary script was showing an unexpected difference
2299        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2300        write a Perl script that can interpret lines of an input file either as
2301        byte characters or as UTF-8, which is what "perltest" was being required to
2302        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2303        can't do is switch easily at run time between having the "use utf8;" pragma
2304        or not. In the end, I fudged it by using the RunPerlTest script to insert
2305        "use utf8;" explicitly for the UTF-8 tests.
2306    
2307     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2308        the end of the subject string, contrary to the documentation and to what
2309        Perl does. This was true of both matching functions. Now it matches only at
2310        the start of the subject and immediately after *internal* newlines.
2311    
2312     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2313        a pointer to an int instead of a pointer to an unsigned long int. This
2314        caused problems on 64-bit systems.
2315    
2316     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2317        instance of the 'standard' template library not being so standard".
2318    
2319    10. There was no check on the number of named subpatterns nor the maximum
2320        length of a subpattern name. The product of these values is used to compute
2321        the size of the memory block for a compiled pattern. By supplying a very
2322        long subpattern name and a large number of named subpatterns, the size
2323        computation could be caused to overflow. This is now prevented by limiting
2324        the length of names to 32 characters, and the number of named subpatterns
2325        to 10,000.
2326    
2327    11. Subpatterns that are repeated with specific counts have to be replicated in
2328        the compiled pattern. The size of memory for this was computed from the
2329        length of the subpattern and the repeat count. The latter is limited to
2330        65535, but there was no limit on the former, meaning that integer overflow
2331        could in principle occur. The compiled length of a repeated subpattern is
2332        now limited to 30,000 bytes in order to prevent this.
2333    
2334    12. Added the optional facility to have named substrings with the same name.
2335    
2336    13. Added the ability to use a named substring as a condition, using the
2337        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2338        are numbers (not recommended). Forward references are permitted.
2339    
2340    14. Added forward references in named backreferences (if you see what I mean).
2341    
2342    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2343        pattern could run off the end of the subject. For example, the pattern
2344        "(?s)(.{1,5})"8 did this with the subject "ab".
2345    
2346    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2347        PCRE_CASELESS was set when matching characters that were quantified with ?
2348        or *.
2349    
2350    17. A character class other than a single negated character that had a minimum
2351        but no maximum quantifier - for example [ab]{6,} - was not handled
2352        correctly by pce_dfa_exec(). It would match only one character.
2353    
2354    18. A valid (though odd) pattern that looked like a POSIX character
2355        class but used an invalid character after [ (for example [[,abc,]]) caused
2356        pcre_compile() to give the error "Failed: internal error: code overflow" or
2357        in some cases to crash with a glibc free() error. This could even happen if
2358        the pattern terminated after [[ but there just happened to be a sequence of
2359        letters, a binary zero, and a closing ] in the memory that followed.
2360    
2361    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2362        over the years. Originally (before any Unicode support), just the bottom 8
2363        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2364        output from "man perlunicode" includes this:
2365    
2366          The regular expression compiler produces polymorphic opcodes.  That
2367          is, the pattern adapts to the data and automatically switches to
2368          the Unicode character scheme when presented with Unicode data--or
2369          instead uses a traditional byte scheme when presented with byte
2370          data.
2371    
2372        Sadly, a wide octal escape does not cause a switch, and in a string with
2373        no other multibyte characters, these octal escapes are treated as before.
2374        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2375        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2376        Unicode string.
2377    
2378        I have not perpetrated such confusion in PCRE. Up till now, it took just
2379        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2380        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2381        translate to the appropriate multibyte character.
2382    
2383    29. Applied some refactoring to reduce the number of warnings from Microsoft
2384        and Borland compilers. This has included removing the fudge introduced
2385        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2386        a warning about an unused variable.
2387    
2388    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2389        characters since release 4.0, because Perl (from release 5.004) does not.
2390        [Or at least, is documented not to: some releases seem to be in conflict
2391        with the documentation.] However, when a pattern was studied with
2392        pcre_study() and all its branches started with \s, PCRE still included VT
2393        as a possible starting character. Of course, this did no harm; it just
2394        caused an unnecessary match attempt.
2395    
2396    22. Removed a now-redundant internal flag bit that recorded the fact that case
2397        dependency changed within the pattern. This was once needed for "required
2398        byte" processing, but is no longer used. This recovers a now-scarce options
2399        bit. Also moved the least significant internal flag bit to the most-
2400        significant bit of the word, which was not previously used (hangover from
2401        the days when it was an int rather than a uint) to free up another bit for
2402        the future.
2403    
2404    23. Added support for CRLF line endings as well as CR and LF. As well as the
2405        default being selectable at build time, it can now be changed at runtime
2406        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2407        specify that it is scanning data with non-default line endings.
2408    
2409    24. Changed the definition of CXXLINK to make it agree with the definition of
2410        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2411    
2412    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2413        recursions. This makes a big different to stack usage for some patterns.
2414    
2415    26. If a subpattern containing a named recursion or subroutine reference such
2416        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2417        the space required for the compiled pattern went wrong and gave too small a
2418        value. Depending on the environment, this could lead to "Failed: internal
2419        error: code overflow at offset 49" or "glibc detected double free or
2420        corruption" errors.
2421    
2422    27. Applied patches from Google (a) to support the new newline modes and (b) to
2423        advance over multibyte UTF-8 characters in GlobalReplace.
2424    
2425    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2426        difference for some implementation of PCRE in some Windows version.
2427    
2428    29. Added some extra testing facilities to pcretest:
2429    
2430        \q<number>   in a data line sets the "match limit" value
2431        \Q<number>   in a data line sets the "match recursion limt" value
2432        -S <number>  sets the stack size, where <number> is in megabytes
2433    
2434        The -S option isn't available for Windows.
2435    
2436    
2437    Version 6.6 06-Feb-06
2438    ---------------------
2439    
2440     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2441        in pcreposix.h. I have copied the definition from pcre.h.
2442    
2443     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2444        because pcre.h is no longer a built file.
2445    
2446     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2447        not normally included in the compiled code.
2448    
2449    
2450  Version 6.5 01-Feb-06  Version 6.5 01-Feb-06
2451  ---------------------  ---------------------
2452    

Legend:
Removed from v.87  
changed lines
  Added in v.758

  ViewVC Help
Powered by ViewVC 1.1.5