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

Diff of /code/trunk/ChangeLog

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

revision 25 by nigel, Sat Feb 24 21:38:45 2007 UTC revision 780 by ph10, Fri Dec 2 11:39:21 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    15. If a forward reference was repeated with an upper limit of around 2000,
59        it caused the error "internal error: overran compiling workspace". The
60        maximum number of forward references (including repeats) was limited by the
61        internal workspace, and dependent on the LINK_SIZE. The code has been
62        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
63        the default depends on LINK_SIZE. There is a new upper limit (for safety)
64        of around 200,000 forward references. While doing this, I also speeded up
65        the filling in of repeated forward references.
66    
67    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
68        incorrectly expecting the subject to contain another "a" after the start.
69    
70    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
71        in the match, the SKIP should be ignored. This was not happening; instead
72        the SKIP was being treated as NOMATCH. For patterns such as
73        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
74        tested.
75    
76    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
77        now much more compatible with Perl, in particular in cases where the result
78        is a non-match for a non-anchored pattern. For example, if
79        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
80        "m", where previously it did not return a name. A side effect of this
81        change is that for partial matches, the last encountered mark name is
82        returned, as for non matches. A number of tests that were previously not
83        Perl-compatible have been moved into the Perl-compatible test files. The
84        refactoring has had the pleasing side effect of removing one argument from
85        the match() function, thus reducing its stack requirements.
86    
87    19. If the /S+ option was used in pcretest to study a pattern using JIT,
88        subsequent uses of /S (without +) incorrectly behaved like /S+.
89    
90    21. Retrieve executable code size support for the JIT compiler and fixing
91        some warnings.
92    
93    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
94        not work when the shorter character appeared right at the end of the
95        subject string.
96    
97    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
98        systems.
99    
100    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
101        output it when the /M option is used in pcretest.
102    
103    
104    Version 8.20 21-Oct-2011
105    ------------------------
106    
107    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
108        a POSIX class. After further experiments with Perl, which convinced me that
109        Perl has bugs and confusions, a closing square bracket is no longer allowed
110        in a POSIX name. This bug also affected patterns with classes that started
111        with full stops.
112    
113    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
114        captured substring, but while checking the failing first alternative,
115        substring 1 is temporarily captured. If the output vector supplied to
116        pcre_exec() was not big enough for this capture, the yield of the function
117        was still zero ("insufficient space for captured substrings"). This cannot
118        be totally fixed without adding another stack variable, which seems a lot
119        of expense for a edge case. However, I have improved the situation in cases
120        such as /(a)(b)x|abc/ matched against "abc", where the return code
121        indicates that fewer than the maximum number of slots in the ovector have
122        been set.
123    
124    3.  Related to (2) above: when there are more back references in a pattern than
125        slots in the output vector, pcre_exec() uses temporary memory during
126        matching, and copies in the captures as far as possible afterwards. It was
127        using the entire output vector, but this conflicts with the specification
128        that only 2/3 is used for passing back captured substrings. Now it uses
129        only the first 2/3, for compatibility. This is, of course, another edge
130        case.
131    
132    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
133        main code base, and can be used by building with --enable-jit. When this is
134        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
135        runtime --no-jit option is given.
136    
137    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
138        ovector, the return from the function was zero, implying that there were
139        other matches that did not fit. The correct "exactly full" value is now
140        returned.
141    
142    6.  If a subpattern that was called recursively or as a subroutine contained
143        (*PRUNE) or any other control that caused it to give a non-standard return,
144        invalid errors such as "Error -26 (nested recursion at the same subject
145        position)" or even infinite loops could occur.
146    
147    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
148        computing the minimum length on reaching *ACCEPT, and so ended up with the
149        wrong value of 1 rather than 0. Further investigation indicates that
150        computing a minimum subject length in the presence of *ACCEPT is difficult
151        (think back references, subroutine calls), and so I have changed the code
152        so that no minimum is registered for a pattern that contains *ACCEPT.
153    
154    8.  If (*THEN) was present in the first (true) branch of a conditional group,
155        it was not handled as intended. [But see 16 below.]
156    
157    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
158        Sheri Pierce.
159    
160    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
161        the first byte in a match must be "a".
162    
163    11. Change 17 for 8.13 increased the recursion depth for patterns like
164        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
165        pattern contains any instances of (*THEN). If it does not, the old
166        optimizations are restored. It would be nice to do this on a per-group
167        basis, but at the moment that is not feasible.
168    
169    12. In some environments, the output of pcretest -C is CRLF terminated. This
170        broke RunTest's code that checks for the link size. A single white space
171        character after the value is now allowed for.
172    
173    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
174        For "fr", it uses the Windows-specific input and output files.
175    
176    14. If (*THEN) appeared in a group that was called recursively or as a
177        subroutine, it did not work as intended. [But see next item.]
178    
179    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
180        pattern fragments (but not containing any | characters). If A and B are
181        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
182        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
183        D. In other words, Perl considers parentheses that do not contain any |
184        characters to be part of a surrounding alternative, whereas PCRE was
185        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
186        differently. PCRE now behaves in the same way as Perl, except in the case
187        of subroutine/recursion calls such as (?1) which have in any case always
188        been different (but PCRE had them first :-).
189    
190    16. Related to 15 above: Perl does not treat the | in a conditional group as
191        creating alternatives. Such a group is treated in the same way as an
192        ordinary group without any | characters when processing (*THEN). PCRE has
193        been changed to match Perl's behaviour.
194    
195    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
196        RunGrepTest script failed.
197    
198    18. Change 22 for version 13 caused atomic groups to use more stack. This is
199        inevitable for groups that contain captures, but it can lead to a lot of
200        stack use in large patterns. The old behaviour has been restored for atomic
201        groups that do not contain any capturing parentheses.
202    
203    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
204        suppress the check for a minimum subject length at run time. (If it was
205        given to pcre_exec() or pcre_dfa_exec() it did work.)
206    
207    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
208        fail to work when decoding hex characters in data strings in EBCDIC
209        environments.
210    
211    21. It appears that in at least one Mac OS environment, the isxdigit() function
212        is implemented as a macro that evaluates to its argument more than once,
213        contravening the C 90 Standard (I haven't checked a later standard). There
214        was an instance in pcretest which caused it to go wrong when processing
215        \x{...} escapes in subject strings. The has been rewritten to avoid using
216        things like p++ in the argument of isxdigit().
217    
218    
219    Version 8.13 16-Aug-2011
220    ------------------------
221    
222    1.  The Unicode data tables have been updated to Unicode 6.0.0.
223    
224    2.  Two minor typos in pcre_internal.h have been fixed.
225    
226    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
227        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
228        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
229    
230    4.  There were a number of related bugs in the code for matching backrefences
231        caselessly in UTF-8 mode when codes for the characters concerned were
232        different numbers of bytes. For example, U+023A and U+2C65 are an upper
233        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
234        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
235        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
236        2-byte code at the end of the subject (it thought there wasn't enough data
237        left).
238    
239    5.  Comprehensive information about what went wrong is now returned by
240        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
241        as the output vector has at least 2 elements. The offset of the start of
242        the failing character and a reason code are placed in the vector.
243    
244    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
245        now returned is for the first byte of the failing character, instead of the
246        last byte inspected. This is an incompatible change, but I hope it is small
247        enough not to be a problem. It makes the returned offset consistent with
248        pcre_exec() and pcre_dfa_exec().
249    
250    7.  pcretest now gives a text phrase as well as the error number when
251        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
252        failure, the offset and reason code are output.
253    
254    8.  When \R was used with a maximizing quantifier it failed to skip backwards
255        over a \r\n pair if the subsequent match failed. Instead, it just skipped
256        back over a single character (\n). This seems wrong (because it treated the
257        two characters as a single entity when going forwards), conflicts with the
258        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
259        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
260        has been changed.
261    
262    9.  Some internal refactoring has changed the processing so that the handling
263        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
264        time (the PCRE_DOTALL option was changed this way some time ago: version
265        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
266        which was always a bit of a fudge. It also means that there is one less
267        argument for the match() function, which reduces its stack requirements
268        slightly. This change also fixes an incompatibility with Perl: the pattern
269        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
270    
271    10. More internal refactoring has drastically reduced the number of recursive
272        calls to match() for possessively repeated groups such as (abc)++ when
273        using pcre_exec().
274    
275    11. While implementing 10, a number of bugs in the handling of groups were
276        discovered and fixed:
277    
278        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
279        (a|)*(?1) gave a compile-time internal error.
280        ((a|)+)+  did not notice that the outer group could match an empty string.
281        (^a|^)+   was not marked as anchored.
282        (.*a|.*)+ was not marked as matching at start or after a newline.
283    
284    12. Yet more internal refactoring has removed another argument from the match()
285        function. Special calls to this function are now indicated by setting a
286        value in a variable in the "match data" data block.
287    
288    13. Be more explicit in pcre_study() instead of relying on "default" for
289        opcodes that mean there is no starting character; this means that when new
290        ones are added and accidentally left out of pcre_study(), testing should
291        pick them up.
292    
293    14. The -s option of pcretest has been documented for ages as being an old
294        synonym of -m (show memory usage). I have changed it to mean "force study
295        for every regex", that is, assume /S for every regex. This is similar to -i
296        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
297        using it. It makes it easier to run collections of tests with and without
298        study enabled, and thereby test pcre_study() more easily. All the standard
299        tests are now run with and without -s (but some patterns can be marked as
300        "never study" - see 20 below).
301    
302    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
303        restoration of the capturing data to the outer values was not happening
304        correctly.
305    
306    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
307        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
308        pattern had matched an empty string, and so incorrectly returned a no
309        match.
310    
311    17. There was optimizing code for the last branch of non-capturing parentheses,
312        and also for the obeyed branch of a conditional subexpression, which used
313        tail recursion to cut down on stack usage. Unfortunately, now that there is
314        the possibility of (*THEN) occurring in these branches, tail recursion is
315        no longer possible because the return has to be checked for (*THEN). These
316        two optimizations have therefore been removed. [But see 8.20/11 above.]
317    
318    18. If a pattern containing \R was studied, it was assumed that \R always
319        matched two bytes, thus causing the minimum subject length to be
320        incorrectly computed because \R can also match just one byte.
321    
322    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
323        was incorrectly computed.
324    
325    20. If /S is present twice on a test pattern in pcretest input, it now
326        *disables* studying, thereby overriding the use of -s on the command line
327        (see 14 above). This is necessary for one or two tests to keep the output
328        identical in both cases.
329    
330    21. When (*ACCEPT) was used in an assertion that matched an empty string and
331        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
332    
333    22. When an atomic group that contained a capturing parenthesis was
334        successfully matched, but the branch in which it appeared failed, the
335        capturing was not being forgotten if a higher numbered group was later
336        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
337        group 1 to "a", when in fact it should be unset. This applied to multi-
338        branched capturing and non-capturing groups, repeated or not, and also to
339        positive assertions (capturing in negative assertions does not happen
340        in PCRE) and also to nested atomic groups.
341    
342    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
343        subject after a captured substring, to make it easier to tell which of a
344        number of identical substrings has been captured.
345    
346    24. The way atomic groups are processed by pcre_exec() has been changed so that
347        if they are repeated, backtracking one repetition now resets captured
348        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
349        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
350        "aaa". Previously, it would have been "a". As part of this code
351        refactoring, the way recursive calls are handled has also been changed.
352    
353    25. If an assertion condition captured any substrings, they were not passed
354        back unless some other capturing happened later. For example, if
355        (?(?=(a))a) was matched against "a", no capturing was returned.
356    
357    26. When studying a pattern that contained subroutine calls or assertions,
358        the code for finding the minimum length of a possible match was handling
359        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
360        group 1 called group 2 while simultaneously a separate group 2 called group
361        1). A stack overflow occurred in this case. I have fixed this by limiting
362        the recursion depth to 10.
363    
364    27. Updated RunTest.bat in the distribution to the version supplied by Tom
365        Fortmann. This supports explicit test numbers on the command line, and has
366        argument validation and error reporting.
367    
368    28. An instance of \X with an unlimited repeat could fail if at any point the
369        first character it looked at was a mark character.
370    
371    29. Some minor code refactoring concerning Unicode properties and scripts
372        should reduce the stack requirement of match() slightly.
373    
374    30. Added the '=' option to pcretest to check the setting of unused capturing
375        slots at the end of the pattern, which are documented as being -1, but are
376        not included in the return count.
377    
378    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
379        compiled something random. Now it gives a compile-time error (as does
380        Perl).
381    
382    32. A *MARK encountered during the processing of a positive assertion is now
383        recorded and passed back (compatible with Perl).
384    
385    33. If --only-matching or --colour was set on a pcregrep call whose pattern
386        had alternative anchored branches, the search for a second match in a line
387        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
388        matched the line "0102" twice. The same bug affected patterns that started
389        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
390        twice.
391    
392    34. Previously, PCRE did not allow quantification of assertions. However, Perl
393        does, and because of capturing effects, quantifying parenthesized
394        assertions may at times be useful. Quantifiers are now allowed for
395        parenthesized assertions.
396    
397    35. A minor code tidy in pcre_compile() when checking options for \R usage.
398    
399    36. \g was being checked for fancy things in a character class, when it should
400        just be a literal "g".
401    
402    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
403        appearance of a nested POSIX class supersedes an apparent external class.
404        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
405        unescaped square brackets may also appear as part of class names. For
406        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
407        more like Perl. (But see 8.20/1 above.)
408    
409    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
410        was because it thought it was \N{name}, which is not supported).
411    
412    39. Add minix to OS list not supporting the -S option in pcretest.
413    
414    40. PCRE tries to detect cases of infinite recursion at compile time, but it
415        cannot analyze patterns in sufficient detail to catch mutual recursions
416        such as ((?1))((?2)). There is now a runtime test that gives an error if a
417        subgroup is called recursively as a subpattern for a second time at the
418        same position in the subject string. In previous releases this might have
419        been caught by the recursion limit, or it might have run out of stack.
420    
421    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
422        happen only once. PCRE was, however incorrectly giving a compile time error
423        "recursive call could loop indefinitely" because it cannot analyze the
424        pattern in sufficient detail. The compile time test no longer happens when
425        PCRE is compiling a conditional subpattern, but actual runaway loops are
426        now caught at runtime (see 40 above).
427    
428    42. It seems that Perl allows any characters other than a closing parenthesis
429        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
430        has been changed to be the same.
431    
432    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
433        as not to get warnings when autogen.sh is called. Also changed
434        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
435    
436    44. To help people who use pcregrep to scan files containing exceedingly long
437        lines, the following changes have been made:
438    
439        (a) The default value of the buffer size parameter has been increased from
440            8K to 20K. (The actual buffer used is three times this size.)
441    
442        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
443            PCRE is built.
444    
445        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
446            to be set at run time.
447    
448        (d) Numerical values in pcregrep options can be followed by K or M, for
449            example --buffer-size=50K.
450    
451        (e) If a line being scanned overflows pcregrep's buffer, an error is now
452            given and the return code is set to 2.
453    
454    45. Add a pointer to the latest mark to the callout data block.
455    
456    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
457        partial match of an empty string instead of no match. This was specific to
458        the use of ".".
459    
460    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
461        complete match instead of a partial match. This bug was dependent on both
462        the PCRE_UTF8 and PCRE_DOTALL options being set.
463    
464    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
465        starting byte set, because \b was not being ignored.
466    
467    
468    Version 8.12 15-Jan-2011
469    ------------------------
470    
471    1.  Fixed some typos in the markup of the man pages, and wrote a script that
472        checks for such things as part of the documentation building process.
473    
474    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
475        --match-limit and --recursion-limit options (added for 8.11). In
476        particular, this made one of the standard tests fail. (The integer value
477        went into the wrong half of a long int.)
478    
479    3.  If the --colour option was given to pcregrep with -v (invert match), it
480        did strange things, either producing crazy output, or crashing. It should,
481        of course, ignore a request for colour when reporting lines that do not
482        match.
483    
484    4.  Another pcregrep bug caused similar problems if --colour was specified with
485        -M (multiline) and the pattern match finished with a line ending.
486    
487    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
488        matched in multiline mode, the following line was shown as part of the
489        match. This seems wrong, so I have changed it.
490    
491    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
492        the check for further matches in the same line (so they could be coloured)
493        to overrun the end of the current line. If another match was found, it was
494        incorrectly shown (and then shown again when found in the next line).
495    
496    7.  If pcregrep was compiled under Windows, there was a reference to the
497        function pcregrep_exit() before it was defined. I am assuming this was
498        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
499        reported by a user. I've moved the definition above the reference.
500    
501    
502    Version 8.11 10-Dec-2010
503    ------------------------
504    
505    1.  (*THEN) was not working properly if there were untried alternatives prior
506        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
507        backtracked to try for "b" instead of moving to the next alternative branch
508        at the same level (in this case, to look for "c"). The Perl documentation
509        is clear that when (*THEN) is backtracked onto, it goes to the "next
510        alternative in the innermost enclosing group".
511    
512    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
513        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
514        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
515        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
516        (*THEN).
517    
518    3.  If \s appeared in a character class, it removed the VT character from
519        the class, even if it had been included by some previous item, for example
520        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
521        of \s, but is part of the POSIX "space" class.)
522    
523    4.  A partial match never returns an empty string (because you can always
524        match an empty string at the end of the subject); however the checking for
525        an empty string was starting at the "start of match" point. This has been
526        changed to the "earliest inspected character" point, because the returned
527        data for a partial match starts at this character. This means that, for
528        example, /(?<=abc)def/ gives a partial match for the subject "abc"
529        (previously it gave "no match").
530    
531    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
532        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
533        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
534        has an implication that the given string is incomplete (because a partial
535        match is preferred over a full match). For this reason, these items now
536        give a partial match in this situation. [Aside: previously, the one case
537        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
538        match rather than a full match, which was wrong by the old rules, but is
539        now correct.]
540    
541    6.  There was a bug in the handling of #-introduced comments, recognized when
542        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
543        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
544        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
545        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
546        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
547        places in pcre_compile().
548    
549    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
550        comments when looking ahead for named forward references to subpatterns,
551        the only newline sequence it recognized was NL. It now handles newlines
552        according to the set newline convention.
553    
554    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
555        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
556        cater for a lack of strerror(). These oversights have been fixed.
557    
558    9.  Added --match-limit and --recursion-limit to pcregrep.
559    
560    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
561    
562    11. When the -o option was used, pcregrep was setting a return code of 1, even
563        when matches were found, and --line-buffered was not being honoured.
564    
565    12. Added an optional parentheses number to the -o and --only-matching options
566        of pcregrep.
567    
568    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
569        can match an empty string. The code to do it in pcretest and pcredemo
570        needed fixing:
571    
572        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
573            only one byte after an empty string match just before CRLF (this case
574            just got forgotten; "any" and "anycrlf" were OK).
575    
576        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
577            mode when an empty string match preceded an ASCII character followed by
578            a non-ASCII character. (The code for advancing by one character rather
579            than one byte was nonsense.)
580    
581        (c) The pcredemo.c sample program did not have any code at all to handle
582            the cases when CRLF is a valid newline sequence.
583    
584    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
585        as a starting offset was within the subject string. There is now a new
586        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
587        negative or greater than the length of the string. In order to test this,
588        pcretest is extended to allow the setting of negative starting offsets.
589    
590    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
591        starting offset points to the beginning of a UTF-8 character was
592        unnecessarily clumsy. I tidied it up.
593    
594    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
595        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
596    
597    17. Nobody had reported that the --include_dir option, which was added in
598        release 7.7 should have been called --include-dir (hyphen, not underscore)
599        for compatibility with GNU grep. I have changed it to --include-dir, but
600        left --include_dir as an undocumented synonym, and the same for
601        --exclude-dir, though that is not available in GNU grep, at least as of
602        release 2.5.4.
603    
604    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
605        characters from a string of bytes) have been redefined so as not to use
606        loops, in order to improve performance in some environments. At the same
607        time, I abstracted some of the common code into auxiliary macros to save
608        repetition (this should not affect the compiled code).
609    
610    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
611        compile-time error is now given if \c is not followed by an ASCII
612        character, that is, a byte less than 128. (In EBCDIC mode, the code is
613        different, and any byte value is allowed.)
614    
615    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
616        START_OPTIMIZE option, which is now allowed at compile time - but just
617        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
618        to pcregrep and other applications that have no direct access to PCRE
619        options. The new /Y option in pcretest sets this option when calling
620        pcre_compile().
621    
622    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
623        back references. Groups containing recursive back references were forced to
624        be atomic by that change, but in the case of named groups, the amount of
625        memory required was incorrectly computed, leading to "Failed: internal
626        error: code overflow". This has been fixed.
627    
628    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
629        pcretest.c, to avoid build problems in some Borland environments.
630    
631    
632    Version 8.10 25-Jun-2010
633    ------------------------
634    
635    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
636        THEN.
637    
638    2.  (*ACCEPT) was not working when inside an atomic group.
639    
640    3.  Inside a character class, \B is treated as a literal by default, but
641        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
642        causes the error). The code is unchanged, but I tidied the documentation.
643    
644    4.  Inside a character class, PCRE always treated \R and \X as literals,
645        whereas Perl faults them if its -w option is set. I have changed PCRE so
646        that it faults them when PCRE_EXTRA is set.
647    
648    5.  Added support for \N, which always matches any character other than
649        newline. (It is the same as "." when PCRE_DOTALL is not set.)
650    
651    6.  When compiling pcregrep with newer versions of gcc which may have
652        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
653        declared with attribute warn_unused_result" were given. Just casting the
654        result to (void) does not stop the warnings; a more elaborate fudge is
655        needed. I've used a macro to implement this.
656    
657    7.  Minor change to pcretest.c to avoid a compiler warning.
658    
659    8.  Added four artifical Unicode properties to help with an option to make
660        \s etc use properties (see next item). The new properties are: Xan
661        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
662    
663    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
664        use Unicode properties. (*UCP) at the start of a pattern can be used to set
665        this option. Modified pcretest to add /W to test this facility. Added
666        REG_UCP to make it available via the POSIX interface.
667    
668    10. Added --line-buffered to pcregrep.
669    
670    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
671        studied, and the match started with a letter with a code point greater than
672        127 whose first byte was different to the first byte of the other case of
673        the letter, the other case of this starting letter was not recognized
674        (#976).
675    
676    12. If a pattern that was studied started with a repeated Unicode property
677        test, for example, \p{Nd}+, there was the theoretical possibility of
678        setting up an incorrect bitmap of starting bytes, but fortunately it could
679        not have actually happened in practice until change 8 above was made (it
680        added property types that matched character-matching opcodes).
681    
682    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
683        possible starting bytes for non-anchored patterns.
684    
685    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
686        \R, and also a number of cases that involve Unicode properties, both
687        explicit and implicit when PCRE_UCP is set.
688    
689    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
690        input, it could crash or give wrong results if characters with values
691        greater than 0xc0 were present in the subject string. (Detail: it assumed
692        UTF-8 input when processing these items.)
693    
694    16. Added a lot of (int) casts to avoid compiler warnings in systems where
695        size_t is 64-bit (#991).
696    
697    17. Added a check for running out of memory when PCRE is compiled with
698        --disable-stack-for-recursion (#990).
699    
700    18. If the last data line in a file for pcretest does not have a newline on
701        the end, a newline was missing in the output.
702    
703    19. The default pcre_chartables.c file recognizes only ASCII characters (values
704        less than 128) in its various bitmaps. However, there is a facility for
705        generating tables according to the current locale when PCRE is compiled. It
706        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
707        characters, are recognized by isspace() and therefore were getting set in
708        these tables, and indeed these tables seem to approximate to ISO 8859. This
709        caused a problem in UTF-8 mode when pcre_study() was used to create a list
710        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
711        which of course cannot start UTF-8 characters. I have changed the code so
712        that only real ASCII characters (less than 128) and the correct starting
713        bytes for UTF-8 encodings are set for characters greater than 127 when in
714        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
715        altogether.)
716    
717    20. Added the /T option to pcretest so as to be able to run tests with non-
718        standard character tables, thus making it possible to include the tests
719        used for 19 above in the standard set of tests.
720    
721    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
722        reference to a subpattern the other side of a comment that contains an
723        opening parenthesis caused either an internal compiling error, or a
724        reference to the wrong subpattern.
725    
726    
727    Version 8.02 19-Mar-2010
728    ------------------------
729    
730    1.  The Unicode data tables have been updated to Unicode 5.2.0.
731    
732    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
733        configured.
734    
735    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
736        original author of that file, following a query about its status.
737    
738    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
739        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
740    
741    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
742        quantifier applied to a forward-referencing subroutine call, could compile
743        incorrect code or give the error "internal error: previously-checked
744        referenced subpattern not found".
745    
746    6.  Both MS Visual Studio and Symbian OS have problems with initializing
747        variables to point to external functions. For these systems, therefore,
748        pcre_malloc etc. are now initialized to local functions that call the
749        relevant global functions.
750    
751    7.  There were two entries missing in the vectors called coptable and poptable
752        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
753        I've fixed the data, and added a kludgy way of testing at compile time that
754        the lengths are correct (equal to the number of opcodes).
755    
756    8.  Following on from 7, I added a similar kludge to check the length of the
757        eint vector in pcreposix.c.
758    
759    9.  Error texts for pcre_compile() are held as one long string to avoid too
760        much relocation at load time. To find a text, the string is searched,
761        counting zeros. There was no check for running off the end of the string,
762        which could happen if a new error number was added without updating the
763        string.
764    
765    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
766    
767    11. \K was not working if it appeared in an atomic group or in a group that
768        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
769        \K is "not well defined" if used in an assertion. PCRE now accepts it if
770        the assertion is positive, but not if it is negative.
771    
772    12. Change 11 fortuitously reduced the size of the stack frame used in the
773        "match()" function of pcre_exec.c by one pointer. Forthcoming
774        implementation of support for (*MARK) will need an extra pointer on the
775        stack; I have reserved it now, so that the stack frame size does not
776        decrease.
777    
778    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
779        item in branch that calls a recursion is a subroutine call - as in the
780        second branch in the above example - was incorrectly given the compile-
781        time error "recursive call could loop indefinitely" because pcre_compile()
782        was not correctly checking the subroutine for matching a non-empty string.
783    
784    14. The checks for overrunning compiling workspace could trigger after an
785        overrun had occurred. This is a "should never occur" error, but it can be
786        triggered by pathological patterns such as hundreds of nested parentheses.
787        The checks now trigger 100 bytes before the end of the workspace.
788    
789    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
790    
791    
792    Version 8.01 19-Jan-2010
793    ------------------------
794    
795    1.  If a pattern contained a conditional subpattern with only one branch (in
796        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
797        computed the wrong minimum data length (which is of course zero for such
798        subpatterns). This could cause incorrect "no match" results.
799    
800    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
801        the pattern is reset in the first branch, pcre_compile() failed with
802        "internal error: code overflow at offset...". This happened only when
803        the reset was to the original external option setting. (An optimization
804        abstracts leading options settings into an external setting, which was the
805        cause of this.)
806    
807    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
808        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
809        assertion pattern did not match (meaning that the assertion was true), it
810        was incorrectly treated as false if the SKIP had been reached during the
811        matching. This also applied to assertions used as conditions.
812    
813    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
814        assertion subpattern, including such a pattern used as a condition,
815        unpredictable results occurred, instead of the error return
816        PCRE_ERROR_DFA_UITEM.
817    
818    5.  The C++ GlobalReplace function was not working like Perl for the special
819        situation when an empty string is matched. It now does the fancy magic
820        stuff that is necessary.
821    
822    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
823        removed. (These were left over from very, very early versions of PCRE.)
824    
825    7.  Some cosmetic changes to the code to make life easier when compiling it
826        as part of something else:
827    
828        (a) Change DEBUG to PCRE_DEBUG.
829    
830        (b) In pcre_compile(), rename the member of the "branch_chain" structure
831            called "current" as "current_branch", to prevent a collision with the
832            Linux macro when compiled as a kernel module.
833    
834        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
835            prevent a collision with the Linux macro when compiled as a kernel
836            module.
837    
838    8.  In pcre_compile() there are some checks for integer overflows that used to
839        cast potentially large values to (double). This has been changed to that
840        when building, a check for int64_t is made, and if it is found, it is used
841        instead, thus avoiding the use of floating point arithmetic. (There is no
842        other use of FP in PCRE.) If int64_t is not found, the fallback is to
843        double.
844    
845    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
846        2005 (difference between two addresses compared to an unsigned value).
847    
848    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
849        custom one, because of the following reported problem in Windows:
850    
851          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
852              under Win32.
853          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
854              therefore missing the function definition.
855          - The compiler thus generates a "C" signature for the test function.
856          - The linker fails to find the "C" function.
857          - PCRE fails to configure if asked to do so against libbz2.
858    
859    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
860        messages were output:
861    
862          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
863          rerunning libtoolize, to keep the correct libtool macros in-tree.
864          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
865    
866        I have done both of these things.
867    
868    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
869        most of the time, it *can* run out if it is given a pattern that contains a
870        runaway infinite recursion. I updated the discussion in the pcrestack man
871        page.
872    
873    13. Now that we have gone to the x.xx style of version numbers, the minor
874        version may start with zero. Using 08 or 09 is a bad idea because users
875        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
876        interpreted as invalid octal numbers. I've updated the previous comment in
877        configure.ac, and also added a check that gives an error if 08 or 09 are
878        used.
879    
880    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
881        causing partial matching to fail when the end of the subject matched \W
882        in a UTF-8 pattern where \W was quantified with a minimum of 3.
883    
884    15. There were some discrepancies between the declarations in pcre_internal.h
885        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
886        their definitions. The declarations used "const uschar *" and the
887        definitions used USPTR. Even though USPTR is normally defined as "const
888        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
889        reported that: "This difference in casting confuses some C++ compilers, for
890        example, SunCC recognizes above declarations as different functions and
891        generates broken code for hbpcre." I have changed the declarations to use
892        USPTR.
893    
894    16. GNU libtool is named differently on some systems. The autogen.sh script now
895        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
896        (FreeBSD).
897    
898    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
899        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
900        comment: "Figure out how to create a longlong from a string: strtoll and
901        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
902        instance, but it only takes 2 args instead of 3!"
903    
904    18. A subtle bug concerned with back references has been fixed by a change of
905        specification, with a corresponding code fix. A pattern such as
906        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
907        refers, was giving matches when it shouldn't. For example, xa=xaaa would
908        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
909        same bug. Such groups have to be quantified to be useful, or contained
910        inside another quantified group. (If there's no repetition, the reference
911        can never match.) The problem arises because, having left the group and
912        moved on to the rest of the pattern, a later failure that backtracks into
913        the group uses the captured value from the final iteration of the group
914        rather than the correct earlier one. I have fixed this in PCRE by forcing
915        any group that contains a reference to itself to be an atomic group; that
916        is, there cannot be any backtracking into it once it has completed. This is
917        similar to recursive and subroutine calls.
918    
919    
920    Version 8.00 19-Oct-09
921    ----------------------
922    
923    1.  The table for translating pcre_compile() error codes into POSIX error codes
924        was out-of-date, and there was no check on the pcre_compile() error code
925        being within the table. This could lead to an OK return being given in
926        error.
927    
928    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
929        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
930        in a Windows environment.
931    
932    3.  The pcregrep --count option prints the count for each file even when it is
933        zero, as does GNU grep. However, pcregrep was also printing all files when
934        --files-with-matches was added. Now, when both options are given, it prints
935        counts only for those files that have at least one match. (GNU grep just
936        prints the file name in this circumstance, but including the count seems
937        more useful - otherwise, why use --count?) Also ensured that the
938        combination -clh just lists non-zero counts, with no names.
939    
940    4.  The long form of the pcregrep -F option was incorrectly implemented as
941        --fixed_strings instead of --fixed-strings. This is an incompatible change,
942        but it seems right to fix it, and I didn't think it was worth preserving
943        the old behaviour.
944    
945    5.  The command line items --regex=pattern and --regexp=pattern were not
946        recognized by pcregrep, which required --regex pattern or --regexp pattern
947        (with a space rather than an '='). The man page documented the '=' forms,
948        which are compatible with GNU grep; these now work.
949    
950    6.  No libpcreposix.pc file was created for pkg-config; there was just
951        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
952    
953    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
954        when UCP support is not needed, by modifying the Python script that
955        generates it from Unicode data files. This should not matter if the module
956        is correctly used as a library, but I received one complaint about 50K of
957        unwanted data. My guess is that the person linked everything into his
958        program rather than using a library. Anyway, it does no harm.
959    
960    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
961        was a minimum greater than 1 for a wide character in a possessive
962        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
963        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
964        character. Chaos in the form of incorrect output or a compiling loop could
965        result.
966    
967    9.  The restrictions on what a pattern can contain when partial matching is
968        requested for pcre_exec() have been removed. All patterns can now be
969        partially matched by this function. In addition, if there are at least two
970        slots in the offset vector, the offset of the earliest inspected character
971        for the match and the offset of the end of the subject are set in them when
972        PCRE_ERROR_PARTIAL is returned.
973    
974    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
975        synonymous with PCRE_PARTIAL, for backwards compatibility, and
976        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
977        and may be more useful for multi-segment matching.
978    
979    11. Partial matching with pcre_exec() is now more intuitive. A partial match
980        used to be given if ever the end of the subject was reached; now it is
981        given only if matching could not proceed because another character was
982        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
983        string "Z", which now yields "no match" instead of "partial match". In the
984        case of pcre_dfa_exec(), "no match" is given if every matching path for the
985        final character ended with (*FAIL).
986    
987    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
988        if the pattern had a "must contain" character that was already found in the
989        earlier partial match, unless partial matching was again requested. For
990        example, with the pattern /dog.(body)?/, the "must contain" character is
991        "g". If the first part-match was for the string "dog", restarting with
992        "sbody" failed. This bug has been fixed.
993    
994    13. The string returned by pcre_dfa_exec() after a partial match has been
995        changed so that it starts at the first inspected character rather than the
996        first character of the match. This makes a difference only if the pattern
997        starts with a lookbehind assertion or \b or \B (\K is not supported by
998        pcre_dfa_exec()). It's an incompatible change, but it makes the two
999        matching functions compatible, and I think it's the right thing to do.
1000    
1001    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1002        so that the demonstration program is easily available in environments where
1003        PCRE has not been installed from source.
1004    
1005    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1006        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1007        library.
1008    
1009    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1010        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1011        is not the first non-POSIX option to be added. Clearly some people find
1012        these options useful.
1013    
1014    17. If a caller to the POSIX matching function regexec() passes a non-zero
1015        value for nmatch with a NULL value for pmatch, the value of
1016        nmatch is forced to zero.
1017    
1018    18. RunGrepTest did not have a test for the availability of the -u option of
1019        the diff command, as RunTest does. It now checks in the same way as
1020        RunTest, and also checks for the -b option.
1021    
1022    19. If an odd number of negated classes containing just a single character
1023        interposed, within parentheses, between a forward reference to a named
1024        subpattern and the definition of the subpattern, compilation crashed with
1025        an internal error, complaining that it could not find the referenced
1026        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1027        [The bug was that it was starting one character too far in when skipping
1028        over the character class, thus treating the ] as data rather than
1029        terminating the class. This meant it could skip too much.]
1030    
1031    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1032        /g option in pcretest when the pattern contains \K, which makes it possible
1033        to have an empty string match not at the start, even when the pattern is
1034        anchored. Updated pcretest and pcredemo to use this option.
1035    
1036    21. If the maximum number of capturing subpatterns in a recursion was greater
1037        than the maximum at the outer level, the higher number was returned, but
1038        with unset values at the outer level. The correct (outer level) value is
1039        now given.
1040    
1041    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1042        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1043        make it do so. The string so far is captured, making this feature
1044        compatible with Perl.
1045    
1046    23. The tests have been re-organized, adding tests 11 and 12, to make it
1047        possible to check the Perl 5.10 features against Perl 5.10.
1048    
1049    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1050        pattern matches a fixed length string. PCRE did not allow this; now it
1051        does. Neither allows recursion.
1052    
1053    25. I finally figured out how to implement a request to provide the minimum
1054        length of subject string that was needed in order to match a given pattern.
1055        (It was back references and recursion that I had previously got hung up
1056        on.) This code has now been added to pcre_study(); it finds a lower bound
1057        to the length of subject needed. It is not necessarily the greatest lower
1058        bound, but using it to avoid searching strings that are too short does give
1059        some useful speed-ups. The value is available to calling programs via
1060        pcre_fullinfo().
1061    
1062    26. While implementing 25, I discovered to my embarrassment that pcretest had
1063        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1064        study optimizations had never been tested with that matching function.
1065        Oops. What is worse, even when it was passed study data, there was a bug in
1066        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1067        were also very few tests of studied patterns with pcre_dfa_exec().
1068    
1069    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1070        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1071        on the other side of the coin, they are no longer allowed to have different
1072        names, because these cannot be distinguished in PCRE, and this has caused
1073        confusion. (This is a difference from Perl.)
1074    
1075    28. When duplicate subpattern names are present (necessarily with different
1076        numbers, as required by 27 above), and a test is made by name in a
1077        conditional pattern, either for a subpattern having been matched, or for
1078        recursion in such a pattern, all the associated numbered subpatterns are
1079        tested, and the overall condition is true if the condition is true for any
1080        one of them. This is the way Perl works, and is also more like the way
1081        testing by number works.
1082    
1083    
1084    Version 7.9 11-Apr-09
1085    ---------------------
1086    
1087    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1088        (pcretest), all targets were linked against these libraries. This included
1089        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1090        libraries. This caused unwanted dependencies to be created. This problem
1091        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1092        pcretest is linked with readline.
1093    
1094    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1095        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1096        moved outside it again, because FALSE and TRUE are already defined in AIX,
1097        but BOOL is not.
1098    
1099    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1100        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1101    
1102    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1103        hyphens) following file names and line numbers when outputting matching
1104        lines. This is not true; no spaces are inserted. I have also clarified the
1105        wording for the --colour (or --color) option.
1106    
1107    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1108        was not coloured; this is different to GNU grep, so I have changed it to be
1109        the same.
1110    
1111    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1112        each matching line was coloured. Now it goes on to look for further matches
1113        of any of the test patterns, which is the same behaviour as GNU grep.
1114    
1115    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1116        doesn't make sense to accept an empty string match in pcregrep, so I have
1117        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1118        seems to be how GNU grep behaves.
1119    
1120    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1121        start or after a newline", because the conditional assertion was not being
1122        correctly handled. The rule now is that both the assertion and what follows
1123        in the first alternative must satisfy the test.
1124    
1125    9.  If auto-callout was enabled in a pattern with a conditional group whose
1126        condition was an assertion, PCRE could crash during matching, both with
1127        pcre_exec() and pcre_dfa_exec().
1128    
1129    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1130        used for matching.
1131    
1132    11. Unicode property support in character classes was not working for
1133        characters (bytes) greater than 127 when not in UTF-8 mode.
1134    
1135    12. Added the -M command line option to pcretest.
1136    
1137    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1138    
1139    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1140    
1141    16. Added comments and documentation about mis-use of no_arg in the C++
1142        wrapper.
1143    
1144    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1145        from Martin Jerabek that uses macro names for all relevant character and
1146        string constants.
1147    
1148    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1149        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1150        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1151        these, but not everybody uses configure.
1152    
1153    19. A conditional group that had only one branch was not being correctly
1154        recognized as an item that could match an empty string. This meant that an
1155        enclosing group might also not be so recognized, causing infinite looping
1156        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1157        with the subject "ab", where knowledge that the repeated group can match
1158        nothing is needed in order to break the loop.
1159    
1160    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1161        exec(), but without supplying a callout function, matching went wrong.
1162    
1163    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1164        leak if the size of the offset vector was greater than 30. When the vector
1165        is smaller, the saved offsets during recursion go onto a local stack
1166        vector, but for larger vectors malloc() is used. It was failing to free
1167        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1168        error, in fact).
1169    
1170    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1171        heapframe that is used only when UTF-8 support is enabled. This caused no
1172        problem, but was untidy.
1173    
1174    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1175        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1176        included within another project.
1177    
1178    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1179        slightly modified by me:
1180    
1181          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1182              not building pcregrep.
1183    
1184          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1185              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1186    
1187    25. Forward references, both numeric and by name, in patterns that made use of
1188        duplicate group numbers, could behave incorrectly or give incorrect errors,
1189        because when scanning forward to find the reference group, PCRE was not
1190        taking into account the duplicate group numbers. A pattern such as
1191        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1192    
1193    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1194        the feature of a custom pointer more persuasive (as requested by a user).
1195    
1196    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1197        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1198        pre-defined.
1199    
1200    28. Added support for (*UTF8) at the start of a pattern.
1201    
1202    29. Arrange for flags added by the "release type" setting in CMake to be shown
1203        in the configuration summary.
1204    
1205    
1206    Version 7.8 05-Sep-08
1207    ---------------------
1208    
1209    1.  Replaced UCP searching code with optimized version as implemented for Ad
1210        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1211        stage table and inline lookup instead of a function, giving speed ups of 2
1212        to 5 times on some simple patterns that I tested. Permission was given to
1213        distribute the MultiStage2.py script that generates the tables (it's not in
1214        the tarball, but is in the Subversion repository).
1215    
1216    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1217        scripts.
1218    
1219    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1220        a group with a zero qualifier. The result of the study could be incorrect,
1221        or the function might crash, depending on the pattern.
1222    
1223    4.  Caseless matching was not working for non-ASCII characters in back
1224        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1225        It now works when Unicode Property Support is available.
1226    
1227    5.  In pcretest, an escape such as \x{de} in the data was always generating
1228        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1229        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1230        truncation.
1231    
1232    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1233    
1234    7.  Added two (int) casts to pcregrep when printing the difference of two
1235        pointers, in case they are 64-bit values.
1236    
1237    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1238        test 2 if it fails.
1239    
1240    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1241        and a #define of that name to empty if it is not externally set. This is to
1242        allow users of MSVC to set it if necessary.
1243    
1244    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1245        the convenience functions in the pcre_get.c source file.
1246    
1247    11. An option change at the start of a pattern that had top-level alternatives
1248        could cause overwriting and/or a crash. This command provoked a crash in
1249        some environments:
1250    
1251          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1252    
1253        This potential security problem was recorded as CVE-2008-2371.
1254    
1255    12. For a pattern where the match had to start at the beginning or immediately
1256        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1257        pcre_dfa_exec() could read past the end of the passed subject if there was
1258        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1259        pcretest so that it places the subject at the end of its malloc-ed buffer.
1260    
1261    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1262        exec() might read past the end of the data buffer in UTF-8 mode.
1263    
1264    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1265        the data contained the byte 0x85 as part of a UTF-8 character within its
1266        first line. This applied both to normal and DFA matching.
1267    
1268    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1269        /^[^d]*?$/8 failed to match "abc".
1270    
1271    16. Added a missing copyright notice to pcrecpp_internal.h.
1272    
1273    17. Make it more clear in the documentation that values returned from
1274        pcre_exec() in ovector are byte offsets, not character counts.
1275    
1276    18. Tidied a few places to stop certain compilers from issuing warnings.
1277    
1278    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1279        supplied by Stefan Weber. I made a further small update for 7.8 because
1280        there is a change of source arrangements: the pcre_searchfuncs.c module is
1281        replaced by pcre_ucd.c.
1282    
1283    
1284    Version 7.7 07-May-08
1285    ---------------------
1286    
1287    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1288        a string to a long long, pretend we don't even have a long long." This is
1289        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1290    
1291    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1292        pre-7.6 versions, which defined a global no_arg variable instead of putting
1293        it in the RE class. (See also #8 below.)
1294    
1295    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1296        Lopes.
1297    
1298    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1299    
1300        (1) The include/exclude patterns were being applied to the whole pathnames
1301            of files, instead of just to the final components.
1302    
1303        (2) If there was more than one level of directory, the subdirectories were
1304            skipped unless they satisfied the include/exclude conditions. This is
1305            inconsistent with GNU grep (and could even be seen as contrary to the
1306            pcregrep specification - which I improved to make it absolutely clear).
1307            The action now is always to scan all levels of directory, and just
1308            apply the include/exclude patterns to regular files.
1309    
1310    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1311        --exclude_dir in the tests to avoid scanning .svn directories.
1312    
1313    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1314        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1315        doesn't support NULs in patterns.
1316    
1317    7.  Added some missing "const"s to declarations of static tables in
1318        pcre_compile.c and pcre_dfa_exec.c.
1319    
1320    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1321        caused by fix #2  above. (Subsequently also a second patch to fix the
1322        first patch. And a third patch - this was a messy problem.)
1323    
1324    9.  Applied Craig's patch to remove the use of push_back().
1325    
1326    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1327        matching function regexec().
1328    
1329    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1330        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1331        references. PCRE supports relative numbers with this syntax (I don't think
1332        Oniguruma does).
1333    
1334    12. Previously, a group with a zero repeat such as (...){0} was completely
1335        omitted from the compiled regex. However, this means that if the group
1336        was called as a subroutine from elsewhere in the pattern, things went wrong
1337        (an internal error was given). Such groups are now left in the compiled
1338        pattern, with a new opcode that causes them to be skipped at execution
1339        time.
1340    
1341    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1342        to the way PCRE behaves:
1343    
1344        (a) A lone ] character is dis-allowed (Perl treats it as data).
1345    
1346        (b) A back reference to an unmatched subpattern matches an empty string
1347            (Perl fails the current match path).
1348    
1349        (c) A data ] in a character class must be notated as \] because if the
1350            first data character in a class is ], it defines an empty class. (In
1351            Perl it is not possible to have an empty class.) The empty class []
1352            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1353            The negative empty class [^] matches any one character, independently
1354            of the DOTALL setting.
1355    
1356    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1357        non-existent subpattern following a character class starting with ']' and
1358        containing () gave an internal compiling error instead of "reference to
1359        non-existent subpattern". Fortunately, when the pattern did exist, the
1360        compiled code was correct. (When scanning forwards to check for the
1361        existencd of the subpattern, it was treating the data ']' as terminating
1362        the class, so got the count wrong. When actually compiling, the reference
1363        was subsequently set up correctly.)
1364    
1365    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1366        it was being rejected as not supported by pcre_dfa_exec(), even though
1367        other assertions are supported. I have made pcre_dfa_exec() support
1368        (*FAIL).
1369    
1370    16. The implementation of 13c above involved the invention of a new opcode,
1371        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1372        cannot be changed at match time, I realized I could make a small
1373        improvement to matching performance by compiling OP_ALLANY instead of
1374        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1375        on the OP_ANY path.
1376    
1377    17. Compiling pcretest on Windows with readline support failed without the
1378        following two fixes: (1) Make the unistd.h include conditional on
1379        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1380    
1381    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1382        ncurses library to be included for pcretest when ReadLine support is
1383        requested, but also to allow for it to be overridden. This patch came from
1384        Daniel Bergström.
1385    
1386    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1387        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1388        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1389        spotting this.
1390    
1391    
1392    Version 7.6 28-Jan-08
1393    ---------------------
1394    
1395    1.  A character class containing a very large number of characters with
1396        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1397        overflow.
1398    
1399    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1400        HAVE_LONG_LONG is not defined.
1401    
1402    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1403        bring them up to date and include new features. This patch includes:
1404    
1405        - Fixed PH's badly added libz and libbz2 support.
1406        - Fixed a problem with static linking.
1407        - Added pcredemo. [But later removed - see 7 below.]
1408        - Fixed dftables problem and added an option.
1409        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1410            HAVE_LONG_LONG.
1411        - Added readline support for pcretest.
1412        - Added an listing of the option settings after cmake has run.
1413    
1414    4.  A user submitted a patch to Makefile that makes it easy to create
1415        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1416        Makefile.am that cause it to include this special target, without
1417        affecting anything else. Note that the same mingw target plus all
1418        the other distribution libraries and programs are now supported
1419        when configuring with CMake (see 6 below) instead of with
1420        Configure/Make.
1421    
1422    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1423        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1424        exported in the Windows port". It has not yet been confirmed that the patch
1425        solves the problem, but it does no harm.
1426    
1427    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1428        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1429        with CMake, and also correct the comment about stack recursion.
1430    
1431    7.  Remove the automatic building of pcredemo from the ./configure system and
1432        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1433        of a program that users should build themselves after PCRE is installed, so
1434        building it automatically is not really right. What is more, it gave
1435        trouble in some build environments.
1436    
1437    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1438    
1439    
1440    Version 7.5 10-Jan-08
1441    ---------------------
1442    
1443    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1444        values in parens when parsing an RE using the C++ wrapper."
1445    
1446    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1447        Characters greater than 255 were excluded from the class instead of being
1448        included.
1449    
1450    3.  The same bug as (2) above applied to negated POSIX classes such as
1451        [:^space:].
1452    
1453    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1454        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1455        I have changed it.
1456    
1457    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1458        first named subpattern) and a construct such as (?&a) would reference the
1459        first named subpattern whose name started with "a" (in other words, the
1460        length check was missing). Both these problems are fixed. "Subpattern name
1461        expected" is now given for (?&) (a zero-length name), and this patch also
1462        makes it give the same error for \k'' (previously it complained that that
1463        was a reference to a non-existent subpattern).
1464    
1465    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1466        this is right because (?- can be followed by option settings as well as by
1467        digits. I have, however, made the messages clearer.
1468    
1469    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1470        than the number used in the conditional) now cause a compile-time error.
1471        This is actually not compatible with Perl, which accepts such patterns, but
1472        treats the conditional as always being FALSE (as PCRE used to), but it
1473        seems to me that giving a diagnostic is better.
1474    
1475    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1476        and messages.
1477    
1478    9.  Fix two occurrences of "backslash" in comments that should have been
1479        "backspace".
1480    
1481    10. Remove two redundant lines of code that can never be obeyed (their function
1482        was moved elsewhere).
1483    
1484    11. The program that makes PCRE's Unicode character property table had a bug
1485        which caused it to generate incorrect table entries for sequences of
1486        characters that have the same character type, but are in different scripts.
1487        It amalgamated them into a single range, with the script of the first of
1488        them. In other words, some characters were in the wrong script. There were
1489        thirteen such cases, affecting characters in the following ranges:
1490    
1491          U+002b0 - U+002c1
1492          U+0060c - U+0060d
1493          U+0061e - U+00612
1494          U+0064b - U+0065e
1495          U+0074d - U+0076d
1496          U+01800 - U+01805
1497          U+01d00 - U+01d77
1498          U+01d9b - U+01dbf
1499          U+0200b - U+0200f
1500          U+030fc - U+030fe
1501          U+03260 - U+0327f
1502          U+0fb46 - U+0fbb1
1503          U+10450 - U+1049d
1504    
1505    12. The -o option (show only the matching part of a line) for pcregrep was not
1506        compatible with GNU grep in that, if there was more than one match in a
1507        line, it showed only the first of them. It now behaves in the same way as
1508        GNU grep.
1509    
1510    13. If the -o and -v options were combined for pcregrep, it printed a blank
1511        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1512        does the same. The return code can be used to tell if there were any
1513        non-matching lines.
1514    
1515    14. Added --file-offsets and --line-offsets to pcregrep.
1516    
1517    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1518        infinitely looping recursion. The bug was that positive lookaheads were not
1519        being skipped when checking for a possible empty match (negative lookaheads
1520        and both kinds of lookbehind were skipped).
1521    
1522    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1523        inclusion of <windows.h> to before rather than after the definition of
1524        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1525    
1526    17. Specifying a possessive quantifier with a specific limit for a Unicode
1527        character property caused pcre_compile() to compile bad code, which led at
1528        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1529        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1530        caused the error; without that there was no problem.
1531    
1532    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1533    
1534    19. Added --enable-pcretest-libreadline.
1535    
1536    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1537        RE::GlobalReplace(). As a result, the number of replacements returned was
1538        double what it should be. I removed one of the increments, but Craig sent a
1539        later patch that removed the other one (the right fix) and added unit tests
1540        that check the return values (which was not done before).
1541    
1542    21. Several CMake things:
1543    
1544        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1545            the names libpcre and libpcreposix, not just pcre and pcreposix.
1546    
1547        (2) The above change means that pcretest and pcregrep are now correctly
1548            linked with the newly-built libraries, not previously installed ones.
1549    
1550        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1551    
1552    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1553        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1554        UTF-8 newline character). The key issue is that the pattern starts .*;
1555        this means that the match must be either at the beginning, or after a
1556        newline. The bug was in the code for advancing after a failed match and
1557        checking that the new position followed a newline. It was not taking
1558        account of UTF-8 characters correctly.
1559    
1560    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1561        character classes. PCRE was not treating the sequence [:...:] as a
1562        character class unless the ... were all letters. Perl, however, seems to
1563        allow any characters between [: and :], though of course it rejects as
1564        unknown any "names" that contain non-letters, because all the known class
1565        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1566        for example, whereas PCRE did not - it did not recognize a POSIX character
1567        class. This seemed a bit dangerous, so the code has been changed to be
1568        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1569        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1570        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1571        Perl does, and where it didn't before.
1572    
1573    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1574        Windows environments %n is disabled by default.
1575    
1576    
1577    Version 7.4 21-Sep-07
1578    ---------------------
1579    
1580    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1581        means that a class such as [\s] counted as "explicit reference to CR or
1582        LF". That isn't really right - the whole point of the change was to try to
1583        help when there was an actual mention of one of the two characters. So now
1584        the change happens only if \r or \n (or a literal CR or LF) character is
1585        encountered.
1586    
1587    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1588        of both had grown to the point where there were only 3 bits left.
1589        Fortunately, there was spare space in the data structure, and so I have
1590        moved the internal flags into a new 16-bit field to free up more option
1591        bits.
1592    
1593    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1594        but did not set the internal JCHANGED flag - either of these is enough to
1595        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1596        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1597        start sets both bits.
1598    
1599    4.  Added options (at build time, compile time, exec time) to change \R from
1600        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1601    
1602    5.  doc/pcresyntax.html was missing from the distribution.
1603    
1604    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1605        compatibility, even though it is no longer used.
1606    
1607    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1608        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1609        windows.h file is present (where different names are used). [This was
1610        reversed later after testing - see 16 below.]
1611    
1612    8.  Changed all #include <config.h> to #include "config.h". There were also
1613        some further <pcre.h> cases that I changed to "pcre.h".
1614    
1615    9.  When pcregrep was used with the --colour option, it missed the line ending
1616        sequence off the lines that it output.
1617    
1618    10. It was pointed out to me that arrays of string pointers cause lots of
1619        relocations when a shared library is dynamically loaded. A technique of
1620        using a single long string with a table of offsets can drastically reduce
1621        these. I have refactored PCRE in four places to do this. The result is
1622        dramatic:
1623    
1624          Originally:                          290
1625          After changing UCP table:            187
1626          After changing error message table:   43
1627          After changing table of "verbs"       36
1628          After changing table of Posix names   22
1629    
1630        Thanks to the folks working on Gregex for glib for this insight.
1631    
1632    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1633        unicode-properties was also set.
1634    
1635    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1636    
1637    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1638        checked only for CRLF.
1639    
1640    14. Added casts to pcretest.c to avoid compiler warnings.
1641    
1642    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1643    
1644    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1645        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1646        entirely. This removes changes made in 7 above.
1647    
1648    17. The CMake files have been updated, and there is now more information about
1649        building with CMake in the NON-UNIX-USE document.
1650    
1651    
1652    Version 7.3 28-Aug-07
1653    ---------------------
1654    
1655     1. In the rejigging of the build system that eventually resulted in 7.1, the
1656        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1657        brackets there is not right, since it causes compilers to look for an
1658        installed pcre.h, not the version that is in the source that is being
1659        compiled (which of course may be different). I have changed it back to:
1660    
1661          #include "pcre.h"
1662    
1663        I have a vague recollection that the change was concerned with compiling in
1664        different directories, but in the new build system, that is taken care of
1665        by the VPATH setting the Makefile.
1666    
1667     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1668        when the subject happened to end in the byte 0x85 (e.g. if the last
1669        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1670        characters but of course it shouldn't be taken as a newline when it is part
1671        of another character. The bug was that, for an unlimited repeat of . in
1672        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1673        characters when looking for a newline.
1674    
1675     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1676    
1677     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1678        in debug output.
1679    
1680     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1681        long printing in the pcrecpp unittest when running under MinGW.
1682    
1683     6. ESC_K was left out of the EBCDIC table.
1684    
1685     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1686        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1687        limit also applies to "virtual nesting" when a pattern is recursive, and in
1688        this case 1000 isn't so big. I have been able to remove this limit at the
1689        expense of backing off one optimization in certain circumstances. Normally,
1690        when pcre_exec() would call its internal match() function recursively and
1691        immediately return the result unconditionally, it uses a "tail recursion"
1692        feature to save stack. However, when a subpattern that can match an empty
1693        string has an unlimited repetition quantifier, it no longer makes this
1694        optimization. That gives it a stack frame in which to save the data for
1695        checking that an empty string has been matched. Previously this was taken
1696        from the 1000-entry workspace that had been reserved. So now there is no
1697        explicit limit, but more stack is used.
1698    
1699     8. Applied Daniel's patches to solve problems with the import/export magic
1700        syntax that is required for Windows, and which was going wrong for the
1701        pcreposix and pcrecpp parts of the library. These were overlooked when this
1702        problem was solved for the main library.
1703    
1704     9. There were some crude static tests to avoid integer overflow when computing
1705        the size of patterns that contain repeated groups with explicit upper
1706        limits. As the maximum quantifier is 65535, the maximum group length was
1707        set at 30,000 so that the product of these two numbers did not overflow a
1708        32-bit integer. However, it turns out that people want to use groups that
1709        are longer than 30,000 bytes (though not repeat them that many times).
1710        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1711        made it possible to implement the integer overflow checks in a much more
1712        dynamic way, which I have now done. The artificial limitation on group
1713        length has been removed - we now have only the limit on the total length of
1714        the compiled pattern, which depends on the LINK_SIZE setting.
1715    
1716    10. Fixed a bug in the documentation for get/copy named substring when
1717        duplicate names are permitted. If none of the named substrings are set, the
1718        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1719        empty string.
1720    
1721    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1722        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1723        because the ] is interpreted as the first data character and the
1724        terminating ] is not found. PCRE has been made compatible with Perl in this
1725        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1726        cause memory overwriting.
1727    
1728    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1729        string has been matched (to stop an infinite loop). It was not recognizing
1730        a conditional subpattern that could match an empty string if that
1731        subpattern was within another subpattern. For example, it looped when
1732        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1733        condition was not nested. This bug has been fixed.
1734    
1735    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1736        past the start of the subject in the presence of bytes with the top bit
1737        set, for example "\x8aBCD".
1738    
1739    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1740        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1741    
1742    14. Optimized (?!) to (*FAIL).
1743    
1744    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1745        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1746        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1747        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1748        does: it's just the validity check that is more restrictive.
1749    
1750    16. Inserted checks for integer overflows during escape sequence (backslash)
1751        processing, and also fixed erroneous offset values for syntax errors during
1752        backslash processing.
1753    
1754    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1755        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1756    
1757    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1758        caused an overrun.
1759    
1760    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1761        something other than just ASCII characters) inside a group that had an
1762        unlimited repeat caused a loop at compile time (while checking to see
1763        whether the group could match an empty string).
1764    
1765    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1766        [\P{Any}] did so. (Error in the code for printing property names.)
1767    
1768    21. An orphan \E inside a character class could cause a crash.
1769    
1770    22. A repeated capturing bracket such as (A)? could cause a wild memory
1771        reference during compilation.
1772    
1773    23. There are several functions in pcre_compile() that scan along a compiled
1774        expression for various reasons (e.g. to see if it's fixed length for look
1775        behind). There were bugs in these functions when a repeated \p or \P was
1776        present in the pattern. These operators have additional parameters compared
1777        with \d, etc, and these were not being taken into account when moving along
1778        the compiled data. Specifically:
1779    
1780        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1781            length.
1782    
1783        (b) An item such as \pL+ within a repeated group could cause crashes or
1784            loops.
1785    
1786        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1787            "reference to non-existent subpattern" error.
1788    
1789        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1790    
1791    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1792        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1793    
1794    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1795    
1796    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1797        character were causing crashes (broken optimization).
1798    
1799    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1800        \p or \P) caused a compile-time loop.
1801    
1802    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1803        break. For example, the unstudied pattern [\r\n]A does not match the string
1804        "\r\nA" because change 7.0/46 below moves the current point on by two
1805        characters after failing to match at the start. However, the pattern \nA
1806        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1807        the same is true. There doesn't seem any very clean way out of this, but
1808        what I have chosen to do makes the common cases work: PCRE now takes note
1809        of whether there can be an explicit match for \r or \n anywhere in the
1810        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1811        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1812        pattern has explicit CR or LF references.
1813    
1814    29. Added (*CR) etc for changing newline setting at start of pattern.
1815    
1816    
1817    Version 7.2 19-Jun-07
1818    ---------------------
1819    
1820     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1821        which is apparently normally available under Windows.
1822    
1823     2. Re-jig the pcregrep tests with different newline settings in an attempt
1824        to make them independent of the local environment's newline setting.
1825    
1826     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1827    
1828     4. Some of the "internals" tests were previously cut out when the link size
1829        was not 2, because the output contained actual offsets. The recent new
1830        "Z" feature of pcretest means that these can be cut out, making the tests
1831        usable with all link sizes.
1832    
1833     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1834        stack recursion. This gives a massive performance boost under BSD, but just
1835        a small improvement under Linux. However, it saves one field in the frame
1836        in all cases.
1837    
1838     6. Added more features from the forthcoming Perl 5.10:
1839    
1840        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1841            recursion call. It refers to the nth most recently opened parentheses.
1842    
1843        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1844            to be opened parentheses.
1845    
1846        (c) Conditions that refer to capturing parentheses can be specified
1847            relatively, for example, (?(-2)... or (?(+3)...
1848    
1849        (d) \K resets the start of the current match so that everything before
1850            is not part of it.
1851    
1852        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1853    
1854        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1855            reference syntax.
1856    
1857        (g) (?| introduces a group in which the numbering of parentheses in each
1858            alternative starts with the same number.
1859    
1860        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1861    
1862     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1863        PCRE_INFO_JCHANGED.
1864    
1865     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1866        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1867        for detecting groups that can match an empty string.
1868    
1869     9. A pattern with a very large number of alternatives (more than several
1870        hundred) was running out of internal workspace during the pre-compile
1871        phase, where pcre_compile() figures out how much memory will be needed. A
1872        bit of new cunning has reduced the workspace needed for groups with
1873        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1874        workspace instead of running out of the 4096 that are available.
1875    
1876    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1877    
1878    11. Applied patch from Google to remove an optimization that didn't quite work.
1879        The report of the bug said:
1880    
1881          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1882          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1883          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1884    
1885    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1886        it matched the wrong number of bytes.
1887    
1888    
1889    Version 7.1 24-Apr-07
1890    ---------------------
1891    
1892     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1893        that is more "standard", making use of automake and other Autotools. There
1894        is some re-arrangement of the files and adjustment of comments consequent
1895        on this.
1896    
1897     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1898        for recursive directory scanning broke on some systems because the files
1899        are not scanned in any specific order and on different systems the order
1900        was different. A call to "sort" has been inserted into RunGrepTest for the
1901        approprate test as a short-term fix. In the longer term there may be an
1902        alternative.
1903    
1904     3. I had an email from Eric Raymond about problems translating some of PCRE's
1905        man pages to HTML (despite the fact that I distribute HTML pages, some
1906        people do their own conversions for various reasons). The problems
1907        concerned the use of low-level troff macros .br and .in. I have therefore
1908        removed all such uses from the man pages (some were redundant, some could
1909        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1910        HTML has been updated to handle .nf/.fi and to complain if it encounters
1911        .br or .in.
1912    
1913     4. Updated comments in configure.ac that get placed in config.h.in and also
1914        arranged for config.h to be included in the distribution, with the name
1915        config.h.generic, for the benefit of those who have to compile without
1916        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1917    
1918     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1919        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1920        makevp.bat for the current PCRE, using the additional files
1921        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1922    
1923     6. A Windows user reported a minor discrepancy with test 2, which turned out
1924        to be caused by a trailing space on an input line that had got lost in his
1925        copy. The trailing space was an accident, so I've just removed it.
1926    
1927     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1928        that is needed.
1929    
1930     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1931        as "const" (a) because they are and (b) because it helps the PHP
1932        maintainers who have recently made a script to detect big data structures
1933        in the php code that should be moved to the .rodata section. I remembered
1934        to update Builducptable as well, so it won't revert if ucptable.h is ever
1935        re-created.
1936    
1937     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1938        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1939        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1940        support is not required. This saves 1.5-2K of code, which is important in
1941        some applications.
1942    
1943        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1944        so as not to refer to the tables, even though these functions will never be
1945        called when UTF-8 support is disabled. Otherwise there are problems with a
1946        shared library.
1947    
1948    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1949    
1950        (a) It was defining its arguments as char * instead of void *.
1951    
1952        (b) It was assuming that all moves were upwards in memory; this was true
1953            a long time ago when I wrote it, but is no longer the case.
1954    
1955        The emulated memove() is provided for those environments that have neither
1956        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1957        is clearly not the case, as these two bugs were recently reported.
1958    
1959    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1960        and Detrail to create the HTML documentation, the .txt form of the man
1961        pages, and it removes trailing spaces from listed files. It also creates
1962        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1963        case, it wraps all the #defines with #ifndefs. This script should be run
1964        before "make dist".
1965    
1966    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1967        with Unicode property support.
1968    
1969        (a) For a maximizing quantifier, if the two different cases of the
1970            character were of different lengths in their UTF-8 codings (there are
1971            some cases like this - I found 11), and the matching function had to
1972            back up over a mixture of the two cases, it incorrectly assumed they
1973            were both the same length.
1974    
1975        (b) When PCRE was configured to use the heap rather than the stack for
1976            recursion during matching, it was not correctly preserving the data for
1977            the other case of a UTF-8 character when checking ahead for a match
1978            while processing a minimizing repeat. If the check also involved
1979            matching a wide character, but failed, corruption could cause an
1980            erroneous result when trying to check for a repeat of the original
1981            character.
1982    
1983    13. Some tidying changes to the testing mechanism:
1984    
1985        (a) The RunTest script now detects the internal link size and whether there
1986            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1987            values substituted by "configure". (The RunGrepTest script already did
1988            this for UTF-8.) The configure.ac script no longer substitutes the
1989            relevant variables.
1990    
1991        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1992            with length and offset values. This means that the output is different
1993            for different internal link sizes. Test 2 is skipped for link sizes
1994            other than 2 because of this, bypassing the problem. Unfortunately,
1995            there was also a test in test 3 (the locale tests) that used /B and
1996            failed for link sizes other than 2. Rather than cut the whole test out,
1997            I have added a new /Z option to pcretest that replaces the length and
1998            offset values with spaces. This is now used to make test 3 independent
1999            of link size. (Test 2 will be tidied up later.)
2000    
2001    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2002        segmentation fault instead of returning the appropriate error message.
2003    
2004    15. In multiline mode when the newline sequence was set to "any", the pattern
2005        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2006        This doesn't seem right; it now treats the CRLF combination as the line
2007        ending, and so does not match in that case. It's only a pattern such as ^$
2008        that would hit this one: something like ^ABC$ would have failed after \r
2009        and then tried again after \r\n.
2010    
2011    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2012        in an attempt to make files that differ only in their line terminators
2013        compare equal. This works on Linux.
2014    
2015    17. Under certain error circumstances pcregrep might try to free random memory
2016        as it exited. This is now fixed, thanks to valgrind.
2017    
2018    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2019        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2020        was because its rules for how to advance for /g after matching an empty
2021        string at the end of a line did not allow for this case. They now check for
2022        it specially.
2023    
2024    20. pcretest is supposed to handle patterns and data of any length, by
2025        extending its buffers when necessary. It was getting this wrong when the
2026        buffer for a data line had to be extended.
2027    
2028    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2029        CRLF as a newline sequence.
2030    
2031    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2032        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2033        I have nevertheless tidied it up.
2034    
2035    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2036    
2037    24. Added a man page for pcre-config.
2038    
2039    
2040    Version 7.0 19-Dec-06
2041    ---------------------
2042    
2043     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2044        moving to gcc 4.1.1.
2045    
2046     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2047        sys/time.h, which is documented as needed for this function. It doesn't
2048        seem to matter on Linux, but it showed up on some releases of OS X.
2049    
2050     3. It seems that there are systems where bytes whose values are greater than
2051        127 match isprint() in the "C" locale. The "C" locale should be the
2052        default when a C program starts up. In most systems, only ASCII printing
2053        characters match isprint(). This difference caused the output from pcretest
2054        to vary, making some of the tests fail. I have changed pcretest so that:
2055    
2056        (a) When it is outputting text in the compiled version of a pattern, bytes
2057            other than 32-126 are always shown as hex escapes.
2058    
2059        (b) When it is outputting text that is a matched part of a subject string,
2060            it does the same, unless a different locale has been set for the match
2061            (using the /L modifier). In this case, it uses isprint() to decide.
2062    
2063     4. Fixed a major bug that caused incorrect computation of the amount of memory
2064        required for a compiled pattern when options that changed within the
2065        pattern affected the logic of the preliminary scan that determines the
2066        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2067        that the computed length was too small. The symptoms of this bug were
2068        either the PCRE error "internal error: code overflow" from pcre_compile(),
2069        or a glibc crash with a message such as "pcretest: free(): invalid next
2070        size (fast)". Examples of patterns that provoked this bug (shown in
2071        pcretest format) are:
2072    
2073          /(?-x: )/x
2074          /(?x)(?-x: \s*#\s*)/
2075          /((?i)[\x{c0}])/8
2076          /(?i:[\x{c0}])/8
2077    
2078        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2079        is now done differently.
2080    
2081     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2082        wrapper classes; (b) implement a new function in the C++ scanner that is
2083        more efficient than the old way of doing things because it avoids levels of
2084        recursion in the regex matching; (c) add a paragraph to the documentation
2085        for the FullMatch() function.
2086    
2087     6. The escape sequence \n was being treated as whatever was defined as
2088        "newline". Not only was this contrary to the documentation, which states
2089        that \n is character 10 (hex 0A), but it also went horribly wrong when
2090        "newline" was defined as CRLF. This has been fixed.
2091    
2092     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2093        was being set to -1 for the "end of line" case (supposedly a value that no
2094        character can have). Though this value is never used (the check for end of
2095        line is "zero bytes in current character"), it caused compiler complaints.
2096        I've changed it to 0xffffffff.
2097    
2098     8. In pcre_version.c, the version string was being built by a sequence of
2099        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2100        string (as it is for production releases) called a macro with an empty
2101        argument. The C standard says the result of this is undefined. The gcc
2102        compiler treats it as an empty string (which was what was wanted) but it is
2103        reported that Visual C gives an error. The source has been hacked around to
2104        avoid this problem.
2105    
2106     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2107        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2108        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2109        of them did).
2110    
2111    10. Originally, pcretest opened its input and output without "b"; then I was
2112        told that "b" was needed in some environments, so it was added for release
2113        5.0 to both the input and output. (It makes no difference on Unix-like
2114        systems.) Later I was told that it is wrong for the input on Windows. I've
2115        now abstracted the modes into two macros, to make it easier to fiddle with
2116        them, and removed "b" from the input mode under Windows.
2117    
2118    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2119    
2120    12. Added -help and --help to pcretest as an official way of being reminded
2121        of the options.
2122    
2123    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2124        and pcrecpp.cc because they annoy compilers at high warning levels.
2125    
2126    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2127    
2128    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2129        scripts are not C programs :-) and which was not noticed because it works
2130        on Linux.
2131    
2132    16. pcretest is supposed to handle any length of pattern and data line (as one
2133        line or as a continued sequence of lines) by extending its input buffer if
2134        necessary. This feature was broken for very long pattern lines, leading to
2135        a string of junk being passed to pcre_compile() if the pattern was longer
2136        than about 50K.
2137    
2138    17. I have done a major re-factoring of the way pcre_compile() computes the
2139        amount of memory needed for a compiled pattern. Previously, there was code
2140        that made a preliminary scan of the pattern in order to do this. That was
2141        OK when PCRE was new, but as the facilities have expanded, it has become
2142        harder and harder to keep it in step with the real compile phase, and there
2143        have been a number of bugs (see for example, 4 above). I have now found a
2144        cunning way of running the real compile function in a "fake" mode that
2145        enables it to compute how much memory it would need, while actually only
2146        ever using a few hundred bytes of working memory and without too many
2147        tests of the mode. This should make future maintenance and development
2148        easier. A side effect of this work is that the limit of 200 on the nesting
2149        depth of parentheses has been removed (though this was never a serious
2150        limitation, I suspect). However, there is a downside: pcre_compile() now
2151        runs more slowly than before (30% or more, depending on the pattern). I
2152        hope this isn't a big issue. There is no effect on runtime performance.
2153    
2154    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2155        newline (only possible for the last line of a file) and it was a
2156        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2157    
2158    19. Added additional timing features to pcretest. (1) The -tm option now times
2159        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2160        separate command line item, by a number that specifies the number of
2161        repeats to use when timing. The default is 50000; this gives better
2162        precision, but takes uncomfortably long for very large patterns.
2163    
2164    20. Extended pcre_study() to be more clever in cases where a branch of a
2165        subpattern has no definite first character. For example, (a*|b*)[cd] would
2166        previously give no result from pcre_study(). Now it recognizes that the
2167        first character must be a, b, c, or d.
2168    
2169    21. There was an incorrect error "recursive call could loop indefinitely" if
2170        a subpattern (or the entire pattern) that was being tested for matching an
2171        empty string contained only one non-empty item after a nested subpattern.
2172        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2173        incorrectly, because the \d was being skipped in the check.
2174    
2175    22. The pcretest program now has a new pattern option /B and a command line
2176        option -b, which is equivalent to adding /B to every pattern. This causes
2177        it to show the compiled bytecode, without the additional information that
2178        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2179        is the same as /B/I).
2180    
2181    23. A new optimization is now able automatically to treat some sequences such
2182        as a*b as a*+b. More specifically, if something simple (such as a character
2183        or a simple class like \d) has an unlimited quantifier, and is followed by
2184        something that cannot possibly match the quantified thing, the quantifier
2185        is automatically "possessified".
2186    
2187    24. A recursive reference to a subpattern whose number was greater than 39
2188        went wrong under certain circumstances in UTF-8 mode. This bug could also
2189        have affected the operation of pcre_study().
2190    
2191    25. Realized that a little bit of performance could be had by replacing
2192        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2193    
2194    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2195    
2196    27. Possessive quantifiers such as a++ were previously implemented by turning
2197        them into atomic groups such as ($>a+). Now they have their own opcodes,
2198        which improves performance. This includes the automatically created ones
2199        from 23 above.
2200    
2201    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2202        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2203        the first matched character to be a colon. This applied both to named and
2204        numbered groups.
2205    
2206    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2207    
2208    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2209        building PCRE on Windows easier, so I have included it in the distribution.
2210    
2211    31. There is now a check in pcretest against a ridiculously large number being
2212        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2213        loop, the loop is abandoned.
2214    
2215    32. Forward references to subpatterns in conditions such as (?(2)...) where
2216        subpattern 2 is defined later cause pcre_compile() to search forwards in
2217        the pattern for the relevant set of parentheses. This search went wrong
2218        when there were unescaped parentheses in a character class, parentheses
2219        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2220    
2221    33. "Subroutine" calls and backreferences were previously restricted to
2222        referencing subpatterns earlier in the regex. This restriction has now
2223        been removed.
2224    
2225    34. Added a number of extra features that are going to be in Perl 5.10. On the
2226        whole, these are just syntactic alternatives for features that PCRE had
2227        previously implemented using the Python syntax or my own invention. The
2228        other formats are all retained for compatibility.
2229    
2230        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2231            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2232            also .NET compatible.
2233    
2234        (b) A recursion or subroutine call to a named group can now be defined as
2235            (?&name) as well as (?P>name).
2236    
2237        (c) A backreference to a named group can now be defined as \k<name> or
2238            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2239            5.10, are also .NET compatible.
2240    
2241        (d) A conditional reference to a named group can now use the syntax
2242            (?(<name>) or (?('name') as well as (?(name).
2243    
2244        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2245            groups (named and numbered) that are never evaluated inline, but can be
2246            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2247            is always false. There may be only one alternative in such a group.
2248    
2249        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2250            as the simple (?(R). The condition is true only if the most recent
2251            recursion is that of the given number or name. It does not search out
2252            through the entire recursion stack.
2253    
2254        (g) The escape \gN or \g{N} has been added, where N is a positive or
2255            negative number, specifying an absolute or relative reference.
2256    
2257    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2258        some "unreachable code" warnings.
2259    
2260    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2261        things, this adds five new scripts.
2262    
2263    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2264        There were also incompatibilities regarding the handling of \Q..\E inside
2265        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2266        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2267    
2268    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2269        matches an empty string, and forcibly breaks the loop. There were bugs in
2270        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2271        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2272        separate and independent bugs (that affected different cases) have been
2273        fixed.
2274    
2275    39. Refactored the code to abolish the use of different opcodes for small
2276        capturing bracket numbers. This is a tidy that I avoided doing when I
2277        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2278        The new approach is not only tidier, it makes it possible to reduce the
2279        memory needed to fix the previous bug (38).
2280    
2281    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2282        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2283        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2284        mode.
2285    
2286    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2287        report.
2288    
2289    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2290        copy construction and assignment in the C++ wrapper.
2291    
2292    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2293        couple of bugs that could have given wrong results in the "--newline=crlf"
2294        case.
2295    
2296    44. Added a number of casts and did some reorganization of signed/unsigned int
2297        variables following suggestions from Dair Grant. Also renamed the variable
2298        "this" as "item" because it is a C++ keyword.
2299    
2300    45. Arranged for dftables to add
2301    
2302          #include "pcre_internal.h"
2303    
2304        to pcre_chartables.c because without it, gcc 4.x may remove the array
2305        definition from the final binary if PCRE is built into a static library and
2306        dead code stripping is activated.
2307    
2308    46. For an unanchored pattern, if a match attempt fails at the start of a
2309        newline sequence, and the newline setting is CRLF or ANY, and the next two
2310        characters are CRLF, advance by two characters instead of one.
2311    
2312    
2313    Version 6.7 04-Jul-06
2314    ---------------------
2315    
2316     1. In order to handle tests when input lines are enormously long, pcretest has
2317        been re-factored so that it automatically extends its buffers when
2318        necessary. The code is crude, but this _is_ just a test program. The
2319        default size has been increased from 32K to 50K.
2320    
2321     2. The code in pcre_study() was using the value of the re argument before
2322        testing it for NULL. (Of course, in any sensible call of the function, it
2323        won't be NULL.)
2324    
2325     3. The memmove() emulation function in pcre_internal.h, which is used on
2326        systems that lack both memmove() and bcopy() - that is, hardly ever -
2327        was missing a "static" storage class specifier.
2328    
2329     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2330        containing an extended class (one that cannot be represented by a bitmap
2331        because it contains high-valued characters or Unicode property items, e.g.
2332        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2333        pattern, but PCRE should not loop if you do not (it no longer does).
2334        [Detail: two cases were found: (a) a repeated subpattern containing an
2335        extended class; (b) a recursive reference to a subpattern that followed a
2336        previous extended class. It wasn't skipping over the extended class
2337        correctly when UTF-8 mode was not set.]
2338    
2339     5. A negated single-character class was not being recognized as fixed-length
2340        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2341        compile error "lookbehind assertion is not fixed length".
2342    
2343     6. The RunPerlTest auxiliary script was showing an unexpected difference
2344        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2345        write a Perl script that can interpret lines of an input file either as
2346        byte characters or as UTF-8, which is what "perltest" was being required to
2347        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2348        can't do is switch easily at run time between having the "use utf8;" pragma
2349        or not. In the end, I fudged it by using the RunPerlTest script to insert
2350        "use utf8;" explicitly for the UTF-8 tests.
2351    
2352     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2353        the end of the subject string, contrary to the documentation and to what
2354        Perl does. This was true of both matching functions. Now it matches only at
2355        the start of the subject and immediately after *internal* newlines.
2356    
2357     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2358        a pointer to an int instead of a pointer to an unsigned long int. This
2359        caused problems on 64-bit systems.
2360    
2361     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2362        instance of the 'standard' template library not being so standard".
2363    
2364    10. There was no check on the number of named subpatterns nor the maximum
2365        length of a subpattern name. The product of these values is used to compute
2366        the size of the memory block for a compiled pattern. By supplying a very
2367        long subpattern name and a large number of named subpatterns, the size
2368        computation could be caused to overflow. This is now prevented by limiting
2369        the length of names to 32 characters, and the number of named subpatterns
2370        to 10,000.
2371    
2372    11. Subpatterns that are repeated with specific counts have to be replicated in
2373        the compiled pattern. The size of memory for this was computed from the
2374        length of the subpattern and the repeat count. The latter is limited to
2375        65535, but there was no limit on the former, meaning that integer overflow
2376        could in principle occur. The compiled length of a repeated subpattern is
2377        now limited to 30,000 bytes in order to prevent this.
2378    
2379    12. Added the optional facility to have named substrings with the same name.
2380    
2381    13. Added the ability to use a named substring as a condition, using the
2382        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2383        are numbers (not recommended). Forward references are permitted.
2384    
2385    14. Added forward references in named backreferences (if you see what I mean).
2386    
2387    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2388        pattern could run off the end of the subject. For example, the pattern
2389        "(?s)(.{1,5})"8 did this with the subject "ab".
2390    
2391    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2392        PCRE_CASELESS was set when matching characters that were quantified with ?
2393        or *.
2394    
2395    17. A character class other than a single negated character that had a minimum
2396        but no maximum quantifier - for example [ab]{6,} - was not handled
2397        correctly by pce_dfa_exec(). It would match only one character.
2398    
2399    18. A valid (though odd) pattern that looked like a POSIX character
2400        class but used an invalid character after [ (for example [[,abc,]]) caused
2401        pcre_compile() to give the error "Failed: internal error: code overflow" or
2402        in some cases to crash with a glibc free() error. This could even happen if
2403        the pattern terminated after [[ but there just happened to be a sequence of
2404        letters, a binary zero, and a closing ] in the memory that followed.
2405    
2406    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2407        over the years. Originally (before any Unicode support), just the bottom 8
2408        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2409        output from "man perlunicode" includes this:
2410    
2411          The regular expression compiler produces polymorphic opcodes.  That
2412          is, the pattern adapts to the data and automatically switches to
2413          the Unicode character scheme when presented with Unicode data--or
2414          instead uses a traditional byte scheme when presented with byte
2415          data.
2416    
2417        Sadly, a wide octal escape does not cause a switch, and in a string with
2418        no other multibyte characters, these octal escapes are treated as before.
2419        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2420        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2421        Unicode string.
2422    
2423        I have not perpetrated such confusion in PCRE. Up till now, it took just
2424        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2425        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2426        translate to the appropriate multibyte character.
2427    
2428    29. Applied some refactoring to reduce the number of warnings from Microsoft
2429        and Borland compilers. This has included removing the fudge introduced
2430        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2431        a warning about an unused variable.
2432    
2433    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2434        characters since release 4.0, because Perl (from release 5.004) does not.
2435        [Or at least, is documented not to: some releases seem to be in conflict
2436        with the documentation.] However, when a pattern was studied with
2437        pcre_study() and all its branches started with \s, PCRE still included VT
2438        as a possible starting character. Of course, this did no harm; it just
2439        caused an unnecessary match attempt.
2440    
2441    22. Removed a now-redundant internal flag bit that recorded the fact that case
2442        dependency changed within the pattern. This was once needed for "required
2443        byte" processing, but is no longer used. This recovers a now-scarce options
2444        bit. Also moved the least significant internal flag bit to the most-
2445        significant bit of the word, which was not previously used (hangover from
2446        the days when it was an int rather than a uint) to free up another bit for
2447        the future.
2448    
2449    23. Added support for CRLF line endings as well as CR and LF. As well as the
2450        default being selectable at build time, it can now be changed at runtime
2451        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2452        specify that it is scanning data with non-default line endings.
2453    
2454    24. Changed the definition of CXXLINK to make it agree with the definition of
2455        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2456    
2457    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2458        recursions. This makes a big different to stack usage for some patterns.
2459    
2460    26. If a subpattern containing a named recursion or subroutine reference such
2461        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2462        the space required for the compiled pattern went wrong and gave too small a
2463        value. Depending on the environment, this could lead to "Failed: internal
2464        error: code overflow at offset 49" or "glibc detected double free or
2465        corruption" errors.
2466    
2467    27. Applied patches from Google (a) to support the new newline modes and (b) to
2468        advance over multibyte UTF-8 characters in GlobalReplace.
2469    
2470    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2471        difference for some implementation of PCRE in some Windows version.
2472    
2473    29. Added some extra testing facilities to pcretest:
2474    
2475        \q<number>   in a data line sets the "match limit" value
2476        \Q<number>   in a data line sets the "match recursion limt" value
2477        -S <number>  sets the stack size, where <number> is in megabytes
2478    
2479        The -S option isn't available for Windows.
2480    
2481    
2482    Version 6.6 06-Feb-06
2483    ---------------------
2484    
2485     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2486        in pcreposix.h. I have copied the definition from pcre.h.
2487    
2488     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2489        because pcre.h is no longer a built file.
2490    
2491     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2492        not normally included in the compiled code.
2493    
2494    
2495    Version 6.5 01-Feb-06
2496    ---------------------
2497    
2498     1. When using the partial match feature with pcre_dfa_exec(), it was not
2499        anchoring the second and subsequent partial matches at the new starting
2500        point. This could lead to incorrect results. For example, with the pattern
2501        /1234/, partially matching against "123" and then "a4" gave a match.
2502    
2503     2. Changes to pcregrep:
2504    
2505        (a) All non-match returns from pcre_exec() were being treated as failures
2506            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2507            error message is output. Some extra information is given for the
2508            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2509            probably the only errors that are likely to be caused by users (by
2510            specifying a regex that has nested indefinite repeats, for instance).
2511            If there are more than 20 of these errors, pcregrep is abandoned.
2512    
2513        (b) A binary zero was treated as data while matching, but terminated the
2514            output line if it was written out. This has been fixed: binary zeroes
2515            are now no different to any other data bytes.
2516    
2517        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2518            used to set a locale for matching. The --locale=xxxx long option has
2519            been added (no short equivalent) to specify a locale explicitly on the
2520            pcregrep command, overriding the environment variables.
2521    
2522        (d) When -B was used with -n, some line numbers in the output were one less
2523            than they should have been.
2524    
2525        (e) Added the -o (--only-matching) option.
2526    
2527        (f) If -A or -C was used with -c (count only), some lines of context were
2528            accidentally printed for the final match.
2529    
2530        (g) Added the -H (--with-filename) option.
2531    
2532        (h) The combination of options -rh failed to suppress file names for files
2533            that were found from directory arguments.
2534    
2535        (i) Added the -D (--devices) and -d (--directories) options.
2536    
2537        (j) Added the -F (--fixed-strings) option.
2538    
2539        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2540    
2541        (l) Added the --colo(u)r option.
2542    
2543        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2544            is not present by default.
2545    
2546     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2547        items such as (?R) or (?1), when the recursion could match a number of
2548        alternatives. If it matched one of the alternatives, but subsequently,
2549        outside the recursion, there was a failure, the code tried to back up into
2550        the recursion. However, because of the way PCRE is implemented, this is not
2551        possible, and the result was an incorrect result from the match.
2552    
2553        In order to prevent this happening, the specification of recursion has
2554        been changed so that all such subpatterns are automatically treated as
2555        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2556    
2557     4. I had overlooked the fact that, in some locales, there are characters for
2558        which isalpha() is true but neither isupper() nor islower() are true. In
2559        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2560        and ordfeminine) are like this. This affected the treatment of \w and \W
2561        when they appeared in character classes, but not when they appeared outside
2562        a character class. The bit map for "word" characters is now created
2563        separately from the results of isalnum() instead of just taking it from the
2564        upper, lower, and digit maps. (Plus the underscore character, of course.)
2565    
2566     5. The above bug also affected the handling of POSIX character classes such as
2567        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2568        permanent tables. Instead, the bit maps for such a class were previously
2569        created as the appropriate unions of the upper, lower, and digit bitmaps.
2570        Now they are created by subtraction from the [[:word:]] class, which has
2571        its own bitmap.
2572    
2573     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2574        It is created by subtracting the vertical space characters (\x09, \x0a,
2575        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2576        subtraction was done in the overall bitmap for a character class, meaning
2577        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2578        be recognized. This bug has been fixed.
2579    
2580     7. Patches from the folks at Google:
2581    
2582          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2583          real life, but is still worth protecting against".
2584    
2585          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2586          regular expressions".
2587    
2588          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2589          have it.
2590    
2591          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2592          "configure" and the latter not, in order to fix a problem somebody had
2593          with compiling the Arg class on HP-UX.
2594    
2595          (e) Improve the error-handling of the C++ wrapper a little bit.
2596    
2597          (f) New tests for checking recursion limiting.
2598    
2599     8. The pcre_memmove() function, which is used only if the environment does not
2600        have a standard memmove() function (and is therefore rarely compiled),
2601        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2602        returning a result (though PCRE never actually uses the result).
2603    
2604     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2605        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2606        returned instead of calling malloc() with an overflowing number that would
2607        most likely cause subsequent chaos.
2608    
2609    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2610    
2611    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2612        with this option is matched, the nmatch and pmatch options of regexec() are
2613        ignored.
2614    
2615    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2616        provided in case anyone wants to the the POSIX interface with UTF-8
2617        strings.
2618    
2619    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2620        C++ linking (needed for some HP-UX environments).
2621    
2622    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2623        (unused parameter) and in the pcre_printint() function (omitted "default"
2624        switch label when the default is to do nothing).
2625    
2626    15. Added some code to make it possible, when PCRE is compiled as a C++
2627        library, to replace subject pointers for pcre_exec() with a smart pointer
2628        class, thus making it possible to process discontinuous strings.
2629    
2630    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2631        much the same function. They were added by different people who were trying
2632        to make PCRE easy to compile on non-Unix systems. It has been suggested
2633        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2634        for compiling on Windows systems. I have therefore replaced it with
2635        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2636        defaults to "extern" for C or "extern C" for C++, which works fine on
2637        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2638        SCOPE with something explicit in config.h. In addition:
2639    
2640        (a) pcreposix.h still had just "extern" instead of either of these macros;
2641            I have replaced it with PCRE_DATA_SCOPE.
2642    
2643        (b) Functions such as _pcre_xclass(), which are internal to the library,
2644            but external in the C sense, all had PCRE_EXPORT in their definitions.
2645            This is apparently wrong for the Windows case, so I have removed it.
2646            (It makes no difference on Unix-like systems.)
2647    
2648    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2649        of recursive calls to match(). This is different to MATCH_LIMIT because
2650        that limits the total number of calls to match(), not all of which increase
2651        the depth of recursion. Limiting the recursion depth limits the amount of
2652        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2653        when PCRE is compiled, and changed at run time. A patch from Google adds
2654        this functionality to the C++ interface.
2655    
2656    18. Changes to the handling of Unicode character properties:
2657    
2658        (a) Updated the table to Unicode 4.1.0.
2659    
2660        (b) Recognize characters that are not in the table as "Cn" (undefined).
2661    
2662        (c) I revised the way the table is implemented to a much improved format
2663            which includes recognition of ranges. It now supports the ranges that
2664            are defined in UnicodeData.txt, and it also amalgamates other
2665            characters into ranges. This has reduced the number of entries in the
2666            table from around 16,000 to around 3,000, thus reducing its size
2667            considerably. I realized I did not need to use a tree structure after
2668            all - a binary chop search is just as efficient. Having reduced the
2669            number of entries, I extended their size from 6 bytes to 8 bytes to
2670            allow for more data.
2671    
2672        (d) Added support for Unicode script names via properties such as \p{Han}.
2673    
2674    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2675        matching that character.
2676    
2677    20. When matching a repeated Unicode property with a minimum greater than zero,
2678        (for example \pL{2,}), PCRE could look past the end of the subject if it
2679        reached it while seeking the minimum number of characters. This could
2680        happen only if some of the characters were more than one byte long, because
2681        there is a check for at least the minimum number of bytes.
2682    
2683    21. Refactored the implementation of \p and \P so as to be more general, to
2684        allow for more different types of property in future. This has changed the
2685        compiled form incompatibly. Anybody with saved compiled patterns that use
2686        \p or \P will have to recompile them.
2687    
2688    22. Added "Any" and "L&" to the supported property types.
2689    
2690    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2691        but give a compile time error if the value is greater than 0xff.
2692    
2693    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2694        accidentally not being installed or uninstalled.
2695    
2696    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2697        made were to insert the current release number. This seemed silly, because
2698        it made things harder for people building PCRE on systems that don't run
2699        "configure". I have turned pcre.h into a distributed file, no longer built
2700        by "configure", with the version identification directly included. There is
2701        no longer a pcre.h.in file.
2702    
2703        However, this change necessitated a change to the pcre-config script as
2704        well. It is built from pcre-config.in, and one of the substitutions was the
2705        release number. I have updated configure.ac so that ./configure now finds
2706        the release number by grepping pcre.h.
2707    
2708    26. Added the ability to run the tests under valgrind.
2709    
2710    
2711    Version 6.4 05-Sep-05
2712    ---------------------
2713    
2714     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2715        "--" to be printed when multiple files were scanned, even when none of the
2716        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2717        consider it to be a bug, and have restored the previous behaviour.
2718    
2719     2. A couple of code tidies to get rid of compiler warnings.
2720    
2721     3. The pcretest program used to cheat by referring to symbols in the library
2722        whose names begin with _pcre_. These are internal symbols that are not
2723        really supposed to be visible externally, and in some environments it is
2724        possible to suppress them. The cheating is now confined to including
2725        certain files from the library's source, which is a bit cleaner.
2726    
2727     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2728        file's purpose clearer.
2729    
2730     5. Reorganized pcre_ucp_findchar().
2731    
2732    
2733    Version 6.3 15-Aug-05
2734    ---------------------
2735    
2736     1. The file libpcre.pc.in did not have general read permission in the tarball.
2737    
2738     2. There were some problems when building without C++ support:
2739    
2740        (a) If C++ support was not built, "make install" and "make test" still
2741            tried to test it.
2742    
2743        (b) There were problems when the value of CXX was explicitly set. Some
2744            changes have been made to try to fix these, and ...
2745    
2746        (c) --disable-cpp can now be used to explicitly disable C++ support.
2747    
2748        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2749            backslash in a target when C++ was disabled. This confuses some
2750            versions of "make", apparently. Using an intermediate variable solves
2751            this. (Same for CPP_LOBJ.)
2752    
2753     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2754        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2755        necessary on certain architectures.
2756    
2757     4. Added a setting of -export-symbols-regex to the link command to remove
2758        those symbols that are exported in the C sense, but actually are local
2759        within the library, and not documented. Their names all begin with
2760        "_pcre_". This is not a perfect job, because (a) we have to except some
2761        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2762        available (and never for static libraries). I have made a note to try to
2763        find a way round (a) in the future.
2764    
2765    
2766    Version 6.2 01-Aug-05
2767    ---------------------
2768    
2769     1. There was no test for integer overflow of quantifier values. A construction
2770        such as {1111111111111111} would give undefined results. What is worse, if
2771        a minimum quantifier for a parenthesized subpattern overflowed and became
2772        negative, the calculation of the memory size went wrong. This could have
2773        led to memory overwriting.
2774    
2775     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2776    
2777     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2778        operating environments where this matters.
2779    
2780     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2781        PCRE options from within the C++ wrapper.
2782    
2783     5. Named capturing subpatterns were not being correctly counted when a pattern
2784        was compiled. This caused two problems: (a) If there were more than 100
2785        such subpatterns, the calculation of the memory needed for the whole
2786        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2787        back references of the form \12, where the number was greater than 9, were
2788        not recognized as back references, even though there were sufficient
2789        previous subpatterns.
2790    
2791     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2792        versions of gcc, e.g. 2.95.4.
2793    
2794    
2795    Version 6.1 21-Jun-05
2796    ---------------------
2797    
2798     1. There was one reference to the variable "posix" in pcretest.c that was not
2799        surrounded by "#if !defined NOPOSIX".
2800    
2801     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2802        the cross-check on the old pcre_info() function, for the benefit of the
2803        cut-down version of PCRE that is currently imported into Exim.
2804    
2805     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2806        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2807        patterns that start (?i) but I don't think that matters. The use of (?i) is
2808        just an example; this all applies to the other options as well.
2809    
2810     4. Since libtool seems to echo the compile commands it is issuing, the output
2811        from "make" can be reduced a bit by putting "@" in front of each libtool
2812        compile command.
2813    
2814     5. Patch from the folks at Google for configure.in to be a bit more thorough
2815        in checking for a suitable C++ installation before trying to compile the
2816        C++ stuff. This should fix a reported problem when a compiler was present,
2817        but no suitable headers.
2818    
2819     6. The man pages all had just "PCRE" as their title. I have changed them to
2820        be the relevant file name. I have also arranged that these names are
2821        retained in the file doc/pcre.txt, which is a concatenation in text format
2822        of all the man pages except the little individual ones for each function.
2823    
2824     7. The NON-UNIX-USE file had not been updated for the different set of source
2825        files that come with release 6. I also added a few comments about the C++
2826        wrapper.
2827    
2828    
2829    Version 6.0 07-Jun-05
2830    ---------------------
2831    
2832     1. Some minor internal re-organization to help with my DFA experiments.
2833    
2834     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2835        didn't matter for the library itself when fully configured, but did matter
2836        when compiling without UCP support, or within Exim, where the ucp files are
2837        not imported.
2838    
2839     3. Refactoring of the library code to split up the various functions into
2840        different source modules. The addition of the new DFA matching code (see
2841        below) to a single monolithic source would have made it really too
2842        unwieldy, quite apart from causing all the code to be include in a
2843        statically linked application, when only some functions are used. This is
2844        relevant even without the DFA addition now that patterns can be compiled in
2845        one application and matched in another.
2846    
2847        The downside of splitting up is that there have to be some external
2848        functions and data tables that are used internally in different modules of
2849        the library but which are not part of the API. These have all had their
2850        names changed to start with "_pcre_" so that they are unlikely to clash
2851        with other external names.
2852    
2853     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2854        a different (DFA) algorithm. Although it is slower than the original
2855        function, it does have some advantages for certain types of matching
2856        problem.
2857    
2858     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2859        including restarting after a partial match.
2860    
2861     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2862        defined when compiling for Windows was sent to me. I have put it into the
2863        code, though I have no means of testing or verifying it.
2864    
2865     7. Added the pcre_refcount() auxiliary function.
2866    
2867     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2868        match before or at the first newline in the subject string. In pcretest,
2869        the /f option on a pattern can be used to set this.
2870    
2871     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2872        would behave wrongly. This has been present in PCRE since release 4.0.
2873    
2874    10. A number of changes to the pcregrep command:
2875    
2876        (a) Refactored how -x works; insert ^(...)$ instead of setting
2877            PCRE_ANCHORED and checking the length, in preparation for adding
2878            something similar for -w.
2879    
2880        (b) Added the -w (match as a word) option.
2881    
2882        (c) Refactored the way lines are read and buffered so as to have more
2883            than one at a time available.
2884    
2885        (d) Implemented a pcregrep test script.
2886    
2887        (e) Added the -M (multiline match) option. This allows patterns to match
2888            over several lines of the subject. The buffering ensures that at least
2889            8K, or the rest of the document (whichever is the shorter) is available
2890            for matching (and similarly the previous 8K for lookbehind assertions).
2891    
2892        (f) Changed the --help output so that it now says
2893    
2894              -w, --word-regex(p)
2895    
2896            instead of two lines, one with "regex" and the other with "regexp"
2897            because that confused at least one person since the short forms are the
2898            same. (This required a bit of code, as the output is generated
2899            automatically from a table. It wasn't just a text change.)
2900    
2901        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2902            option but starts with a hyphen. Could be a pattern or a path name
2903            starting with a hyphen, for instance.
2904    
2905        (h) "-" can be given as a file name to represent stdin.
2906    
2907        (i) When file names are being printed, "(standard input)" is used for
2908            the standard input, for compatibility with GNU grep. Previously
2909            "<stdin>" was used.
2910    
2911        (j) The option --label=xxx can be used to supply a name to be used for
2912            stdin when file names are being printed. There is no short form.
2913    
2914        (k) Re-factored the options decoding logic because we are going to add
2915            two more options that take data. Such options can now be given in four
2916            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2917    
2918        (l) Added the -A, -B, and -C options for requesting that lines of context
2919            around matches be printed.
2920    
2921        (m) Added the -L option to print the names of files that do not contain
2922            any matching lines, that is, the complement of -l.
2923    
2924        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2925            continue to scan other files.
2926    
2927        (o) The -s option was incorrectly implemented. For compatibility with other
2928            greps, it now suppresses the error message for a non-existent or non-
2929            accessible file (but not the return code). There is a new option called
2930            -q that suppresses the output of matching lines, which was what -s was
2931            previously doing.
2932    
2933        (p) Added --include and --exclude options to specify files for inclusion
2934            and exclusion when recursing.
2935    
2936    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2937        Hopefully, it now does.
2938    
2939    12. Missing cast in pcre_study().
2940    
2941    13. Added an "uninstall" target to the makefile.
2942    
2943    14. Replaced "extern" in the function prototypes in Makefile.in with
2944        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2945        world, but is set differently for Windows.
2946    
2947    15. Added a second compiling function called pcre_compile2(). The only
2948        difference is that it has an extra argument, which is a pointer to an
2949        integer error code. When there is a compile-time failure, this is set
2950        non-zero, in addition to the error test pointer being set to point to an
2951        error message. The new argument may be NULL if no error number is required
2952        (but then you may as well call pcre_compile(), which is now just a
2953        wrapper). This facility is provided because some applications need a
2954        numeric error indication, but it has also enabled me to tidy up the way
2955        compile-time errors are handled in the POSIX wrapper.
2956    
2957    16. Added VPATH=.libs to the makefile; this should help when building with one
2958        prefix path and installing with another. (Or so I'm told by someone who
2959        knows more about this stuff than I do.)
2960    
2961    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2962        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2963        match everything, including newlines. This is not POSIX-compatible, but
2964        somebody wanted the feature. From pcretest it can be activated by using
2965        both the P and the s flags.
2966    
2967    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2968    
2969    19. libpcre.pc was being incorrectly installed as executable.
2970    
2971    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2972        it now also looks for '\r' so that it will work unmodified on Windows.
2973    
2974    21. Added Google's contributed C++ wrapper to the distribution.
2975    
2976    22. Added some untidy missing memory free() calls in pcretest, to keep
2977        Electric Fence happy when testing.
2978    
2979    
2980    
2981    Version 5.0 13-Sep-04
2982    ---------------------
2983    
2984     1. Internal change: literal characters are no longer packed up into items
2985        containing multiple characters in a single byte-string. Each character
2986        is now matched using a separate opcode. However, there may be more than one
2987        byte in the character in UTF-8 mode.
2988    
2989     2. The pcre_callout_block structure has two new fields: pattern_position and
2990        next_item_length. These contain the offset in the pattern to the next match
2991        item, and its length, respectively.
2992    
2993     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2994        insertion of callouts before each pattern item. Added the /C option to
2995        pcretest to make use of this.
2996    
2997     4. On the advice of a Windows user, the lines
2998    
2999          #if defined(_WIN32) || defined(WIN32)
3000          _setmode( _fileno( stdout ), 0x8000 );
3001          #endif  /* defined(_WIN32) || defined(WIN32) */
3002    
3003        have been added to the source of pcretest. This apparently does useful
3004        magic in relation to line terminators.
3005    
3006     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3007        for the benefit of those environments where the "b" makes a difference.
3008    
3009     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3010        to know about it. I have put a hack into configure.in that adds in code
3011        to set GCC=yes if CC=icc. This seems to end up at a point in the
3012        generated configure script that is early enough to affect the setting of
3013        compiler options, which is what is needed, but I have no means of testing
3014        whether it really works. (The user who reported this had patched the
3015        generated configure script, which of course I cannot do.)
3016    
3017        LATER: After change 22 below (new libtool files), the configure script
3018        seems to know about icc (and also ecc). Therefore, I have commented out
3019        this hack in configure.in.
3020    
3021     7. Added support for pkg-config (2 patches were sent in).
3022    
3023     8. Negated POSIX character classes that used a combination of internal tables
3024        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3025        [[:^ascii]]. Typically, they would match almost any characters. The other
3026        POSIX classes were not broken in this way.
3027    
3028     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3029        to find the match, as PCRE was deluded into thinking that the match had to
3030        start at the start point or following a newline. The same bug applied to
3031        patterns with negative forward assertions or any backward assertions
3032        preceding ".*" at the start, unless the pattern required a fixed first
3033        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3034    
3035    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3036        starting at the last subject character, bytes beyond the end of the subject
3037        string were read.
3038    
3039    11. Renamed the variable "class" as "classbits" to make life easier for C++
3040        users. (Previously there was a macro definition, but it apparently wasn't
3041        enough.)
3042    
3043    12. Added the new field "tables" to the extra data so that tables can be passed
3044        in at exec time, or the internal tables can be re-selected. This allows
3045        a compiled regex to be saved and re-used at a later time by a different
3046        program that might have everything at different addresses.
3047    
3048    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3049        -R library as well as a -L library.
3050    
3051    14. The debugging options of pcretest (-d on the command line or D on a
3052        pattern) showed incorrect output for anything following an extended class
3053        that contained multibyte characters and which was followed by a quantifier.
3054    
3055    15. Added optional support for general category Unicode character properties
3056        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3057        support. It adds about 90K to the size of the library. The meanings of the
3058        inbuilt class escapes such as \d and \s have NOT been changed.
3059    
3060    16. Updated pcredemo.c to include calls to free() to release the memory for the
3061        compiled pattern.
3062    
3063    17. The generated file chartables.c was being created in the source directory
3064        instead of in the building directory. This caused the build to fail if the
3065        source directory was different from the building directory, and was
3066        read-only.
3067    
3068    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3069        file. No doubt somebody will tell me if they don't make sense... Also added
3070        Dan Mooney's comments about building on OpenVMS.
3071    
3072    19. Added support for partial matching via the PCRE_PARTIAL option for
3073        pcre_exec() and the \P data escape in pcretest.
3074    
3075    20. Extended pcretest with 3 new pattern features:
3076    
3077        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3078              write the compiled pattern to the file whose name is "rest-of-line".
3079              This is a straight binary dump of the data, with the saved pointer to
3080              the character tables forced to be NULL. The study data, if any, is
3081              written too. After writing, pcretest reads a new pattern.
3082    
3083        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3084              compiled pattern from the given file. There must not be any
3085              occurrences of "<" in the file name (pretty unlikely); if there are,
3086              pcretest will instead treat the initial "<" as a pattern delimiter.
3087              After reading in the pattern, pcretest goes on to read data lines as
3088              usual.
3089    
3090        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3091              and 16-bit fields in a compiled pattern, to simulate a pattern that
3092              was compiled on a host of opposite endianness.
3093    
3094    21. The pcre-exec() function can now cope with patterns that were compiled on
3095        hosts of opposite endianness, with this restriction:
3096    
3097          As for any compiled expression that is saved and used later, the tables
3098          pointer field cannot be preserved; the extra_data field in the arguments
3099          to pcre_exec() should be used to pass in a tables address if a value
3100          other than the default internal tables were used at compile time.
3101    
3102    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3103        now diagnosed as an error. Previously, most of the time, a negative number
3104        would have been treated as zero, but if in addition "ovector" was passed as
3105        NULL, a crash could occur.
3106    
3107    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3108        new versions from the libtool 1.5 distribution (the last one is a copy of
3109        a file called libtool.m4). This seems to have fixed the need to patch
3110        "configure" to support Darwin 1.3 (which I used to do). However, I still
3111        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3112        workstation).
3113    
3114    24. Changed the PCRE licence to be the more standard "BSD" licence.
3115    
3116    
3117    Version 4.5 01-Dec-03
3118    ---------------------
3119    
3120     1. There has been some re-arrangement of the code for the match() function so
3121        that it can be compiled in a version that does not call itself recursively.
3122        Instead, it keeps those local variables that need separate instances for
3123        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3124        needs to "recurse". Keeping track of where control must go is done by means
3125        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3126        hide most of the details from the main code, and operates only if
3127        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3128        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3129        operating.
3130    
3131        To make it easier for callers to provide specially tailored get/free
3132        functions for this usage, two new functions, pcre_stack_malloc, and
3133        pcre_stack_free, are used. They are always called in strict stacking order,
3134        and the size of block requested is always the same.
3135    
3136        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3137        PCRE has been compiled to use the stack or the heap for recursion. The
3138        -C option of pcretest uses this to show which version is compiled.
3139    
3140        A new data escape \S, is added to pcretest; it causes the amounts of store
3141        obtained and freed by both kinds of malloc/free at match time to be added
3142        to the output.
3143    
3144     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3145        what's available on my current Linux desktop machine.
3146    
3147     3. When matching a UTF-8 string, the test for a valid string at the start has
3148        been extended. If start_offset is not zero, PCRE now checks that it points
3149        to a byte that is the start of a UTF-8 character. If not, it returns
3150        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3151        this is necessary because there may be backward assertions in the pattern.
3152        When matching the same subject several times, it may save resources to use
3153        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3154    
3155     4. The code for checking the validity of UTF-8 strings has been tightened so
3156        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3157        containing "overlong sequences".
3158    
3159     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3160        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3161        should have been "&", but it just so happened that all the cases this let
3162        through by mistake were picked up later in the function.
3163    
3164     6. I had used a variable called "isblank" - this is a C99 function, causing
3165        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3166    
3167     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3168        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3169        so the version is shown; (c) stop "make test" echoing "./RunTest".
3170    
3171     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3172    
3173     9. The prototype for memmove() for systems that don't have it was using
3174        size_t, but the inclusion of the header that defines size_t was later. I've
3175        moved the #includes for the C headers earlier to avoid this.
3176    
3177    10. Added some adjustments to the code to make it easier to compiler on certain
3178        special systems:
3179    
3180          (a) Some "const" qualifiers were missing.
3181          (b) Added the macro EXPORT before all exported functions; by default this
3182              is defined to be empty.
3183          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3184              that it reads its output file name as an argument instead of writing
3185              to the standard output and assuming this can be redirected.
3186    
3187    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3188        class containing characters with values greater than 255, PCRE compilation
3189        went into a loop.
3190    
3191    12. A recursive reference to a subpattern that was within another subpattern
3192        that had a minimum quantifier of zero caused PCRE to crash. For example,
3193        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3194        recursion. If the recursively-called subpattern itself had a zero repeat,
3195        that was OK.
3196    
3197    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3198        buffer into which it was copied (for escape processing) was still set at
3199        1024, so long lines caused crashes.
3200    
3201    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3202        "internal error: code overflow...". This applied to any character class
3203        that was followed by a possessive quantifier.
3204    
3205    15. Modified the Makefile to add libpcre.la as a prerequisite for
3206        libpcreposix.la because I was told this is needed for a parallel build to
3207        work.
3208    
3209    16. If a pattern that contained .* following optional items at the start was
3210        studied, the wrong optimizing data was generated, leading to matching
3211        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3212        matching string must start with a or b or c. The correct conclusion for
3213        this pattern is that a match can start with any character.
3214    
3215    
3216    Version 4.4 13-Aug-03
3217    ---------------------
3218    
3219     1. In UTF-8 mode, a character class containing characters with values between
3220        127 and 255 was not handled correctly if the compiled pattern was studied.
3221        In fixing this, I have also improved the studying algorithm for such
3222        classes (slightly).
3223    
3224     2. Three internal functions had redundant arguments passed to them. Removal
3225        might give a very teeny performance improvement.
3226    
3227     3. Documentation bug: the value of the capture_top field in a callout is *one
3228        more than* the number of the hightest numbered captured substring.
3229    
3230     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3231        in incorrectly linking with a previously installed version. They now link
3232        explicitly with libpcre.la.
3233    
3234     5. configure.in no longer needs to recognize Cygwin specially.
3235    
3236     6. A problem in pcre.in for Windows platforms is fixed.
3237    
3238     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3239        pcretest, it used to include the size of the study block as part of its
3240        output. Unfortunately, the structure contains a field that has a different
3241        size on different hardware architectures. This meant that the tests that
3242        showed this size failed. As the block is currently always of a fixed size,
3243        this information isn't actually particularly useful in pcretest output, so
3244        I have just removed it.
3245    
3246     8. Three pre-processor statements accidentally did not start in column 1.
3247        Sadly, there are *still* compilers around that complain, even though
3248        standard C has not required this for well over a decade. Sigh.
3249    
3250     9. In pcretest, the code for checking callouts passed small integers in the
3251        callout_data field, which is a void * field. However, some picky compilers
3252        complained about the casts involved for this on 64-bit systems. Now
3253        pcretest passes the address of the small integer instead, which should get
3254        rid of the warnings.
3255    
3256    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3257        both compile and run time, and gives an error if an invalid UTF-8 sequence
3258        is found. There is a option for disabling this check in cases where the
3259        string is known to be correct and/or the maximum performance is wanted.
3260    
3261    11. In response to a bug report, I changed one line in Makefile.in from
3262    
3263            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3264        to
3265            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3266    
3267        to look similar to other lines, but I have no way of telling whether this
3268        is the right thing to do, as I do not use Windows. No doubt I'll get told
3269        if it's wrong...
3270    
3271    
3272    Version 4.3 21-May-03
3273    ---------------------
3274    
3275    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3276       Makefile.
3277    
3278    2. Some refactoring to improve the quality of the code:
3279    
3280       (i)   The utf8_table... variables are now declared "const".
3281    
3282       (ii)  The code for \cx, which used the "case flipping" table to upper case
3283             lower case letters, now just substracts 32. This is ASCII-specific,
3284             but the whole concept of \cx is ASCII-specific, so it seems
3285             reasonable.
3286    
3287       (iii) PCRE was using its character types table to recognize decimal and
3288             hexadecimal digits in the pattern. This is silly, because it handles
3289             only 0-9, a-f, and A-F, but the character types table is locale-
3290             specific, which means strange things might happen. A private
3291             table is now used for this - though it costs 256 bytes, a table is
3292             much faster than multiple explicit tests. Of course, the standard
3293             character types table is still used for matching digits in subject
3294             strings against \d.
3295    
3296       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3297             ending in _t are). So I've renamed it as ESC_tee.
3298    
3299    3. The first argument for regexec() in the POSIX wrapper should have been
3300       defined as "const".
3301    
3302    4. Changed pcretest to use malloc() for its buffers so that they can be
3303       Electric Fenced for debugging.
3304    
3305    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3306       to read one or more bytes before the start of the subject string. Often this
3307       had no effect on PCRE's behaviour, but in some circumstances it could
3308       provoke a segmentation fault.
3309    
3310    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3311       to try to read one or more bytes before the start of the subject string.
3312    
3313    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3314       UTF-8 support could misbehave in various ways if the subject string
3315       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3316       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3317       back over UTF-8 characters.)
3318    
3319    
3320    Version 4.2 14-Apr-03
3321    ---------------------
3322    
3323    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3324    
3325    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3326         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3327         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3328         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3329         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3330           and BUILD_EXEEXT
3331         Note: automatic setting of the BUILD variables is not yet working
3332         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3333           compile-time but not at link-time
3334         [LINK]: use for linking executables only
3335         make different versions for Windows and non-Windows
3336         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3337           libraries
3338         [LINK_FOR_BUILD]: new variable
3339         [OBJEXT]: use throughout
3340         [EXEEXT]: use throughout
3341         <winshared>: new target
3342         <wininstall>: new target
3343         <dftables.o>: use native compiler
3344         <dftables>: use native linker
3345         <install>: handle Windows platform correctly
3346         <clean>: ditto
3347         <check>: ditto
3348         copy DLL to top builddir before testing
3349    
3350       As part of these changes, -no-undefined was removed again. This was reported
3351       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3352       in any case.
3353    
3354    3. Some tidies to get rid of compiler warnings:
3355    
3356       . In the match_data structure, match_limit was an unsigned long int, whereas
3357         match_call_count was an int. I've made them both unsigned long ints.
3358    
3359       . In pcretest the fact that a const uschar * doesn't automatically cast to
3360         a void * provoked a warning.
3361    
3362       . Turning on some more compiler warnings threw up some "shadow" variables
3363         and a few more missing casts.
3364    
3365    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3366       option, a class that contained a single character with a value between 128
3367       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3368    
3369    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3370       option, a class that contained several characters, but with at least one
3371       whose value was between 128 and 255 caused PCRE to crash.
3372    
3373    
3374    Version 4.1 12-Mar-03
3375    ---------------------
3376    
3377    1. Compiling with gcc -pedantic found a couple of places where casts were
3378    needed, and a string in dftables.c that was longer than standard compilers are
3379    required to support.
3380    
3381    2. Compiling with Sun's compiler found a few more places where the code could
3382    be tidied up in order to avoid warnings.
3383    
3384    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3385    first of these names is deprecated in the latest Autoconf in favour of the name
3386    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3387    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3388    analogy I have changed it to CFLAGS_FOR_BUILD.
3389    
3390    4. Added -no-undefined to the linking command in the Makefile, because this is
3391    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3392    linking step for the pcreposix library.
3393    
3394    5. PCRE was failing to diagnose the case of two named groups with the same
3395    name.
3396    
3397    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3398    literal character that is needed in the subject for a match, and scans along to
3399    ensure that it is present before embarking on the full matching process. This
3400    saves time in cases of nested unlimited repeats that are never going to match.
3401    Problem: the scan can take a lot of time if the subject is very long (e.g.
3402    megabytes), thus penalizing straightforward matches. It is now done only if the
3403    amount of subject to be scanned is less than 1000 bytes.
3404    
3405    7. A lesser problem with the same optimization is that it was recording the
3406    first character of an anchored pattern as "needed", thus provoking a search
3407    right along the subject, even when the first match of the pattern was going to
3408    fail. The "needed" character is now not set for anchored patterns, unless it
3409    follows something in the pattern that is of non-fixed length. Thus, it still
3410    fulfils its original purpose of finding quick non-matches in cases of nested
3411    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3412    
3413    
3414    Version 4.0 17-Feb-03
3415    ---------------------
3416    
3417    1. If a comment in an extended regex that started immediately after a meta-item
3418    extended to the end of string, PCRE compiled incorrect data. This could lead to
3419    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3420    
3421    2. Moved to autoconf 2.53 and libtool 1.4.2.
3422    
3423    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3424    the special perltest8 script is no longer needed - all the tests can be run
3425    from a single perltest script.
3426    
3427    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3428    by \s. It has now been removed in PCRE. This means it isn't recognized as
3429    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3430    class [:space:] *does* include VT, thereby creating a mess.
3431    
3432    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3433    space and tab.
3434    
3435    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3436    its new features into the main test script, reducing the number of scripts.
3437    
3438    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3439    were backward compatible, and made the (?i) apply to the whole pattern, as if
3440    /i were given. Now it behaves more logically, and applies the option setting
3441    only to what follows. PCRE has been changed to follow suit. However, if it
3442    finds options settings right at the start of the pattern, it extracts them into
3443    the global options, as before. Thus, they show up in the info data.
3444    
3445    8. Added support for the \Q...\E escape sequence. Characters in between are
3446    treated as literals. This is slightly different from Perl in that $ and @ are
3447    also handled as literals inside the quotes. In Perl, they will cause variable
3448    interpolation. Note the following examples:
3449    
3450        Pattern            PCRE matches      Perl matches
3451    
3452        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3453        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3454        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3455    
3456    For compatibility with Perl, \Q...\E sequences are recognized inside character
3457    classes as well as outside them.
3458    
3459    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3460    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3461    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3462    signed/unsigned warnings.
3463    
3464    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3465    option for pcretest, so I've replaced it by a simple function that does just
3466    that job.
3467    
3468    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3469    "pcregrep -".
3470    
3471    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3472    Java package. This provides some syntactic sugar for simple cases of what my
3473    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3474    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3475    item, you can use this simplified notation. Note that only makes sense with
3476    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3477    greediness, whatever the setting of the PCRE_UNGREEDY option.
3478    
3479    13. A change of greediness default within a pattern was not taking effect at
3480    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3481    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3482    was abstracted outside.
3483    
3484    14. PCRE now supports the \G assertion. It is true when the current matching
3485    position is at the start point of the match. This differs from \A when the
3486    starting offset is non-zero. Used with the /g option of pcretest (or similar
3487    code), it works in the same way as it does for Perl's /g option. If all
3488    alternatives of a regex begin with \G, the expression is anchored to the start
3489    match position, and the "anchored" flag is set in the compiled expression.
3490    
3491    15. Some bugs concerning the handling of certain option changes within patterns
3492    have been fixed. These applied to options other than (?ims). For example,
3493    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3494    been the other way round. Some of this was related to change 7 above.
3495    
3496    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3497    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3498    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3499    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3500    
3501    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3502    mode. Unlike ".", it always matches newline, whatever the setting of
3503    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3504    assertions. Perl allows it, but it doesn't (in general) work because it can't
3505    calculate the length of the lookbehind. At least, that's the case for Perl
3506    5.8.0 - I've been told they are going to document that it doesn't work in
3507    future.
3508    
3509    18. Added an error diagnosis for escapes that PCRE does not support: these are
3510    \L, \l, \N, \P, \p, \U, \u, and \X.
3511    
3512    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3513    reading past the end of the pattern in cases such as /[abcd/.
3514    
3515    20. PCRE was getting more memory than necessary for patterns with classes that
3516    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3517    
3518    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3519    compiling PCRE for use with Virtual Pascal.
3520    
3521    22. Small fix to the Makefile to make it work properly if the build is done
3522    outside the source tree.
3523    
3524    23. Added a new extension: a condition to go with recursion. If a conditional
3525    subpattern starts with (?(R) the "true" branch is used if recursion has
3526    happened, whereas the "false" branch is used only at the top level.
3527    
3528    24. When there was a very long string of literal characters (over 255 bytes
3529    without UTF support, over 250 bytes with UTF support), the computation of how
3530    much memory was required could be incorrect, leading to segfaults or other
3531    strange effects.
3532    
3533    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3534    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3535    there was a subsequent back reference to those brackets. This meant that, for
3536    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3537    possible to check for precisely this case. All we can do is abandon the
3538    optimization if .* occurs inside capturing brackets when there are any back
3539    references whatsoever. (See below for a better fix that came later.)
3540    
3541    26. The handling of the optimization for finding the first character of a
3542    non-anchored pattern, and for finding a character that is required later in the
3543    match were failing in some cases. This didn't break the matching; it just
3544    failed to optimize when it could. The way this is done has been re-implemented.
3545    
3546    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3547    
3548    28. Added a new feature that provides some of the functionality that Perl
3549    provides with (?{...}). The facility is termed a "callout". The way it is done
3550    in PCRE is for the caller to provide an optional function, by setting
3551    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3552    global variable. By default it is unset, which disables all calling out. To get
3553    the function called, the regex must include (?C) at appropriate points. This
3554    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3555    This provides a means of identifying different callout points. When PCRE
3556    reaches such a point in the regex, if pcre_callout has been set, the external
3557    function is called. It is provided with data in a structure called
3558    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3559    matching continues; if it returns a non-zero value, the match at the current
3560    point fails. However, backtracking will occur if possible. [This was changed
3561    later and other features added - see item 49 below.]
3562    
3563    29. pcretest is upgraded to test the callout functionality. It provides a
3564    callout function that displays information. By default, it shows the start of
3565    the match and the current position in the text. There are some new data escapes
3566    to vary what happens:
3567    
3568        \C+         in addition, show current contents of captured substrings
3569        \C-         do not supply a callout function
3570        \C!n        return 1 when callout number n is reached
3571        \C!n!m      return 1 when callout number n is reached for the mth time
3572    
3573    30. If pcregrep was called with the -l option and just a single file name, it
3574    output "<stdin>" if a match was found, instead of the file name.
3575    
3576    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3577    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3578    pcre_exec(). This saves a malloc/free per call. The default value of
3579    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3580    when configuring.
3581    
3582    32. The default maximum size of a compiled pattern is 64K. There have been a
3583    few cases of people hitting this limit. The code now uses macros to handle the
3584    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3585    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3586    configuring. Tests 2 and 5 work only with 2-byte links because they output
3587    debugging information about compiled patterns.
3588    
3589    33. Internal code re-arrangements:
3590    
3591    (a) Moved the debugging function for printing out a compiled regex into
3592        its own source file (printint.c) and used #include to pull it into
3593        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3594        separate copies.
3595    
3596    (b) Defined the list of op-code names for debugging as a macro in
3597        internal.h so that it is next to the definition of the opcodes.
3598    
3599    (c) Defined a table of op-code lengths for simpler skipping along compiled
3600        code. This is again a macro in internal.h so that it is next to the
3601        definition of the opcodes.
3602    
3603    34. Added support for recursive calls to individual subpatterns, along the
3604    lines of Robin Houston's patch (but implemented somewhat differently).
3605    
3606    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3607    allow it to read and process whole directories in Win32. This code was
3608    contributed by Lionel Fourquaux; it has not been tested by me.
3609    
3610    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3611    used to name a group. Names consist of alphanumerics and underscores, and must
3612    be unique. Back references use the syntax (?P=name) and recursive calls use
3613    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3614    numbers. The function pcre_fullinfo() can be used after compilation to extract
3615    a name/number map. There are three relevant calls:
3616    
3617      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3618      PCRE_INFO_NAMECOUNT            yields the number of entries
3619      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3620    
3621    The map is a vector of fixed-size entries. The size of each entry depends on
3622    the length of the longest name used. The first two bytes of each entry are the
3623    group number, most significant byte first. There follows the corresponding
3624    name, zero terminated. The names are in alphabetical order.
3625    
3626    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3627    case instead of 255. Making it the same both with and without UTF-8 support
3628    means that the same test output works with both.
3629    
3630    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3631    calling malloc() with a zero argument.
3632    
3633    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3634    optimization. I've improved things by keeping a bitmap of backreferences with
3635    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3636    fact referenced, the optimization can be applied. It is unlikely that a
3637    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3638    the match to follow \n) will appear inside brackets with a number greater than
3639    31, but if it does, any back reference > 31 suppresses the optimization.
3640    
3641    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3642    of disabling numbered capturing parentheses. Any opening parenthesis that is
3643    not followed by ? behaves as if it were followed by ?: but named parentheses
3644    can still be used for capturing (and they will acquire numbers in the usual
3645    way).
3646    
3647    41. Redesigned the return codes from the match() function into yes/no/error so
3648    that errors can be passed back from deep inside the nested calls. A malloc
3649    failure while inside a recursive subpattern call now causes the
3650    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3651    
3652    42. It is now possible to set a limit on the number of times the match()
3653    function is called in a call to pcre_exec(). This facility makes it possible to
3654    limit the amount of recursion and backtracking, though not in a directly
3655    obvious way, because the match() function is used in a number of different
3656    circumstances. The count starts from zero for each position in the subject
3657    string (for non-anchored patterns). The default limit is, for compatibility, a
3658    large number, namely 10 000 000. You can change this in two ways:
3659    
3660    (a) When configuring PCRE before making, you can use --with-match-limit=n
3661        to set a default value for the compiled library.
3662    
3663    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3664        a different value is set. See 45 below.
3665    
3666    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3667    
3668    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3669    of things that can be changed at compile time. The first argument specifies
3670    what is wanted and the second points to where the information is to be placed.
3671    The current list of available information is:
3672    
3673      PCRE_CONFIG_UTF8
3674    
3675    The output is an integer that is set to one if UTF-8 support is available;
3676    otherwise it is set to zero.
3677    
3678      PCRE_CONFIG_NEWLINE
3679    
3680    The output is an integer that it set to the value of the code that is used for
3681    newline. It is either LF (10) or CR (13).
3682    
3683      PCRE_CONFIG_LINK_SIZE
3684    
3685    The output is an integer that contains the number of bytes used for internal
3686    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3687    
3688      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3689    
3690    The output is an integer that contains the threshold above which the POSIX
3691    interface uses malloc() for output vectors. See item 31 above.
3692    
3693      PCRE_CONFIG_MATCH_LIMIT
3694    
3695    The output is an unsigned integer that contains the default limit of the number
3696    of match() calls in a pcre_exec() execution. See 42 above.
3697    
3698    44. pcretest has been upgraded by the addition of the -C option. This causes it
3699    to extract all the available output from the new pcre_config() function, and to
3700    output it. The program then exits immediately.
3701    
3702    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3703    order to support additional features. One way would have been to define
3704    pcre_exec2() (for example) with extra arguments, but this would not have been
3705    extensible, and would also have required all calls to the original function to
3706    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3707    is used for passing in "extra" data from pcre_study().
3708    
3709    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3710    contains the following fields:
3711    
3712      flags         a bitmap indicating which of the following fields are set
3713      study_data    opaque data from pcre_study()
3714      match_limit   a way of specifying a limit on match() calls for a specific
3715                      call to pcre_exec()
3716      callout_data  data for callouts (see 49 below)
3717    
3718    The flag bits are also defined in pcre.h, and are
3719    
3720      PCRE_EXTRA_STUDY_DATA
3721      PCRE_EXTRA_MATCH_LIMIT
3722      PCRE_EXTRA_CALLOUT_DATA
3723    
3724    The pcre_study() function now returns one of these new pcre_extra blocks, with
3725    the actual study data pointed to by the study_data field, and the
3726    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3727    before. That is, this change is entirely upwards-compatible and requires no
3728    change to existing code.
3729    
3730    If you want to pass in additional data to pcre_exec(), you can either place it
3731    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3732    block.
3733    
3734    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3735    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3736    times with different match limits, until it finds the minimum value needed for
3737    pcre_exec() to complete. The value is then output. This can be instructive; for
3738    most simple matches the number is quite small, but for pathological cases it
3739    gets very large very quickly.
3740    
3741    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3742    returns the size of the data block pointed to by the study_data field in a
3743    pcre_extra block, that is, the value that was passed as the argument to
3744    pcre_malloc() when PCRE was getting memory in which to place the information
3745    created by pcre_study(). The fourth argument should point to a size_t variable.
3746    pcretest has been extended so that this information is shown after a successful
3747    pcre_study() call when information about the compiled regex is being displayed.
3748    
3749    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3750    because what follows is always an absolute path. (Later: it turns out that this
3751    is more than cosmetic for MinGW, because it doesn't like empty path
3752    components.)
3753    
3754    49. Some changes have been made to the callout feature (see 28 above):
3755    
3756    (i)  A callout function now has three choices for what it returns:
3757    
3758           0  =>  success, carry on matching
3759         > 0  =>  failure at this point, but backtrack if possible
3760         < 0  =>  serious error, return this value from pcre_exec()
3761    
3762         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3763         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3764         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3765         use by callout functions. It will never be used by PCRE itself.
3766    
3767    (ii) The pcre_extra structure (see 45 above) has a void * field called
3768         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3769         pcre_callout_block structure has a field of the same name. The contents of
3770         the field passed in the pcre_extra structure are passed to the callout
3771         function in the corresponding field in the callout block. This makes it
3772         easier to use the same callout-containing regex from multiple threads. For
3773         testing, the pcretest program has a new data escape
3774    
3775           \C*n        pass the number n (may be negative) as callout_data
3776    
3777         If the callout function in pcretest receives a non-zero value as
3778         callout_data, it returns that value.
3779    
3780    50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3781    there were some redundant $(CFLAGS) in commands that are now specified as
3782    $(LINK), which already includes $(CFLAGS).
3783    
3784    51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3785    has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3786    with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3787    one-byte characters throughout. Note that case-insensitive matching applies
3788    only to characters whose values are less than 256. PCRE doesn't support the
3789    notion of cases for higher-valued characters.
3790    
3791    (i)   A character class whose characters are all within 0-255 is handled as
3792          a bit map, and the map is inverted for negative classes. Previously, a
3793          character > 255 always failed to match such a class; however it should
3794          match if the class was a negative one (e.g. [^ab]). This has been fixed.