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

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 889 by ph10, Tue Jan 17 14:51:50 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.30
5    ------------
6    
7    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
8        name is defined in ctype.h.
9    
10    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
11        only in quite long subpatterns.
12    
13    3.  Removed the function pcre_info(), which has been obsolete and deprecated
14        since it was replaced by pcre_fullinfo() in February 2000.
15    
16    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
17        match a (*MARK), and the match failed at the start of the subject, a
18        reference to memory before the start of the subject could occur. This bug
19        was introduced by fix 17 of release 8.21.
20    
21    5.  A reference to an unset group with zero minimum repetition was giving
22        totally wrong answers (in non-JavaScript-compatibility mode). For example,
23        /(another)?(\1?)test/ matched against "hello world test". This bug was
24        introduced in release 8.13.
25    
26    6.  Add support for 16-bit character strings (a large amount of work involving
27        many changes and refactorings).
28    
29    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
30        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
31        from a file.
32    
33    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
34        rounding is not applied in this particular case).
35    
36    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
37        if they appear, or are escaped, in patterns.
38    
39    10. Get rid of a number of -Wunused-but-set-variable warnings.
40    
41    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
42        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
43        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
44        also returns the mark "x". This bug applied to capturing parentheses,
45        non-capturing parentheses, and atomic parentheses. It also applied to some
46        assertions.
47    
48    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
49        information out of configure.ac instead of relying on pcre.h.generic, which
50        is not stored in the repository.
51    
52    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
53        -lreadline.
54    
55    
56    Version 8.21 12-Dec-2011
57    ------------------------
58    
59    1.  Updating the JIT compiler.
60    
61    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
62        are added as well.
63    
64    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
65        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
66        calling _pcre_jit_exec. Some extra comments are added.
67    
68    4.  (*MARK) settings inside atomic groups that do not contain any capturing
69        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
70        was introduced by change 18 for 8.20.
71    
72    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
73        ECMA-262 standard.
74    
75    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
76        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
77        This bug was probably introduced by change 9 of 8.13.
78    
79    7.  While fixing 6 above, I noticed that a number of other items were being
80        incorrectly rejected as "not fixed length". This arose partly because newer
81        opcodes had not been added to the fixed-length checking code. I have (a)
82        corrected the bug and added tests for these items, and (b) arranged for an
83        error to occur if an unknown opcode is encountered while checking for fixed
84        length instead of just assuming "not fixed length". The items that were
85        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
86        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
87        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
88    
89    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
90        being incorrectly compiled and would have given unpredicatble results.
91    
92    9.  A possessively repeated subpattern with minimum repeat count greater than
93        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
94        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
95        the first (A) could occur when it should not.
96    
97    10. Add a cast and remove a redundant test from the code.
98    
99    11. JIT should use pcre_malloc/pcre_free for allocation.
100    
101    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
102        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
103        is anything other than /usr, -L is still shown).
104    
105    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
106    
107    14. Perl does not support \N without a following name in a [] class; PCRE now
108        also gives an error.
109    
110    15. If a forward reference was repeated with an upper limit of around 2000,
111        it caused the error "internal error: overran compiling workspace". The
112        maximum number of forward references (including repeats) was limited by the
113        internal workspace, and dependent on the LINK_SIZE. The code has been
114        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
115        the default depends on LINK_SIZE. There is a new upper limit (for safety)
116        of around 200,000 forward references. While doing this, I also speeded up
117        the filling in of repeated forward references.
118    
119    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
120        incorrectly expecting the subject to contain another "a" after the start.
121    
122    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
123        in the match, the SKIP should be ignored. This was not happening; instead
124        the SKIP was being treated as NOMATCH. For patterns such as
125        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
126        tested.
127    
128    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
129        now much more compatible with Perl, in particular in cases where the result
130        is a non-match for a non-anchored pattern. For example, if
131        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
132        "m", where previously it did not return a name. A side effect of this
133        change is that for partial matches, the last encountered mark name is
134        returned, as for non matches. A number of tests that were previously not
135        Perl-compatible have been moved into the Perl-compatible test files. The
136        refactoring has had the pleasing side effect of removing one argument from
137        the match() function, thus reducing its stack requirements.
138    
139    19. If the /S+ option was used in pcretest to study a pattern using JIT,
140        subsequent uses of /S (without +) incorrectly behaved like /S+.
141    
142    21. Retrieve executable code size support for the JIT compiler and fixing
143        some warnings.
144    
145    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
146        not work when the shorter character appeared right at the end of the
147        subject string.
148    
149    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
150        systems.
151    
152    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
153        output it when the /M option is used in pcretest.
154    
155    25. The CheckMan script was not being included in the distribution. Also, added
156        an explicit "perl" to run Perl scripts from the PrepareRelease script
157        because this is reportedly needed in Windows.
158    
159    26. If study data was being save in a file and studying had not found a set of
160        "starts with" bytes for the pattern, the data written to the file (though
161        never used) was taken from uninitialized memory and so caused valgrind to
162        complain.
163    
164    27. Updated RunTest.bat as provided by Sheri Pierce.
165    
166    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
167    
168    29. Computation of memory usage for the table of capturing group names was
169        giving an unnecessarily large value.
170    
171    
172    Version 8.20 21-Oct-2011
173    ------------------------
174    
175    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
176        a POSIX class. After further experiments with Perl, which convinced me that
177        Perl has bugs and confusions, a closing square bracket is no longer allowed
178        in a POSIX name. This bug also affected patterns with classes that started
179        with full stops.
180    
181    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
182        captured substring, but while checking the failing first alternative,
183        substring 1 is temporarily captured. If the output vector supplied to
184        pcre_exec() was not big enough for this capture, the yield of the function
185        was still zero ("insufficient space for captured substrings"). This cannot
186        be totally fixed without adding another stack variable, which seems a lot
187        of expense for a edge case. However, I have improved the situation in cases
188        such as /(a)(b)x|abc/ matched against "abc", where the return code
189        indicates that fewer than the maximum number of slots in the ovector have
190        been set.
191    
192    3.  Related to (2) above: when there are more back references in a pattern than
193        slots in the output vector, pcre_exec() uses temporary memory during
194        matching, and copies in the captures as far as possible afterwards. It was
195        using the entire output vector, but this conflicts with the specification
196        that only 2/3 is used for passing back captured substrings. Now it uses
197        only the first 2/3, for compatibility. This is, of course, another edge
198        case.
199    
200    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
201        main code base, and can be used by building with --enable-jit. When this is
202        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
203        runtime --no-jit option is given.
204    
205    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
206        ovector, the return from the function was zero, implying that there were
207        other matches that did not fit. The correct "exactly full" value is now
208        returned.
209    
210    6.  If a subpattern that was called recursively or as a subroutine contained
211        (*PRUNE) or any other control that caused it to give a non-standard return,
212        invalid errors such as "Error -26 (nested recursion at the same subject
213        position)" or even infinite loops could occur.
214    
215    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
216        computing the minimum length on reaching *ACCEPT, and so ended up with the
217        wrong value of 1 rather than 0. Further investigation indicates that
218        computing a minimum subject length in the presence of *ACCEPT is difficult
219        (think back references, subroutine calls), and so I have changed the code
220        so that no minimum is registered for a pattern that contains *ACCEPT.
221    
222    8.  If (*THEN) was present in the first (true) branch of a conditional group,
223        it was not handled as intended. [But see 16 below.]
224    
225    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
226        Sheri Pierce.
227    
228    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
229        the first byte in a match must be "a".
230    
231    11. Change 17 for 8.13 increased the recursion depth for patterns like
232        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
233        pattern contains any instances of (*THEN). If it does not, the old
234        optimizations are restored. It would be nice to do this on a per-group
235        basis, but at the moment that is not feasible.
236    
237    12. In some environments, the output of pcretest -C is CRLF terminated. This
238        broke RunTest's code that checks for the link size. A single white space
239        character after the value is now allowed for.
240    
241    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
242        For "fr", it uses the Windows-specific input and output files.
243    
244    14. If (*THEN) appeared in a group that was called recursively or as a
245        subroutine, it did not work as intended. [But see next item.]
246    
247    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
248        pattern fragments (but not containing any | characters). If A and B are
249        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
250        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
251        D. In other words, Perl considers parentheses that do not contain any |
252        characters to be part of a surrounding alternative, whereas PCRE was
253        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
254        differently. PCRE now behaves in the same way as Perl, except in the case
255        of subroutine/recursion calls such as (?1) which have in any case always
256        been different (but PCRE had them first :-).
257    
258    16. Related to 15 above: Perl does not treat the | in a conditional group as
259        creating alternatives. Such a group is treated in the same way as an
260        ordinary group without any | characters when processing (*THEN). PCRE has
261        been changed to match Perl's behaviour.
262    
263    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
264        RunGrepTest script failed.
265    
266    18. Change 22 for version 13 caused atomic groups to use more stack. This is
267        inevitable for groups that contain captures, but it can lead to a lot of
268        stack use in large patterns. The old behaviour has been restored for atomic
269        groups that do not contain any capturing parentheses.
270    
271    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
272        suppress the check for a minimum subject length at run time. (If it was
273        given to pcre_exec() or pcre_dfa_exec() it did work.)
274    
275    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
276        fail to work when decoding hex characters in data strings in EBCDIC
277        environments.
278    
279    21. It appears that in at least one Mac OS environment, the isxdigit() function
280        is implemented as a macro that evaluates to its argument more than once,
281        contravening the C 90 Standard (I haven't checked a later standard). There
282        was an instance in pcretest which caused it to go wrong when processing
283        \x{...} escapes in subject strings. The has been rewritten to avoid using
284        things like p++ in the argument of isxdigit().
285    
286    
287    Version 8.13 16-Aug-2011
288    ------------------------
289    
290    1.  The Unicode data tables have been updated to Unicode 6.0.0.
291    
292    2.  Two minor typos in pcre_internal.h have been fixed.
293    
294    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
295        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
296        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
297    
298    4.  There were a number of related bugs in the code for matching backrefences
299        caselessly in UTF-8 mode when codes for the characters concerned were
300        different numbers of bytes. For example, U+023A and U+2C65 are an upper
301        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
302        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
303        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
304        2-byte code at the end of the subject (it thought there wasn't enough data
305        left).
306    
307    5.  Comprehensive information about what went wrong is now returned by
308        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
309        as the output vector has at least 2 elements. The offset of the start of
310        the failing character and a reason code are placed in the vector.
311    
312    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
313        now returned is for the first byte of the failing character, instead of the
314        last byte inspected. This is an incompatible change, but I hope it is small
315        enough not to be a problem. It makes the returned offset consistent with
316        pcre_exec() and pcre_dfa_exec().
317    
318    7.  pcretest now gives a text phrase as well as the error number when
319        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
320        failure, the offset and reason code are output.
321    
322    8.  When \R was used with a maximizing quantifier it failed to skip backwards
323        over a \r\n pair if the subsequent match failed. Instead, it just skipped
324        back over a single character (\n). This seems wrong (because it treated the
325        two characters as a single entity when going forwards), conflicts with the
326        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
327        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
328        has been changed.
329    
330    9.  Some internal refactoring has changed the processing so that the handling
331        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
332        time (the PCRE_DOTALL option was changed this way some time ago: version
333        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
334        which was always a bit of a fudge. It also means that there is one less
335        argument for the match() function, which reduces its stack requirements
336        slightly. This change also fixes an incompatibility with Perl: the pattern
337        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
338    
339    10. More internal refactoring has drastically reduced the number of recursive
340        calls to match() for possessively repeated groups such as (abc)++ when
341        using pcre_exec().
342    
343    11. While implementing 10, a number of bugs in the handling of groups were
344        discovered and fixed:
345    
346        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
347        (a|)*(?1) gave a compile-time internal error.
348        ((a|)+)+  did not notice that the outer group could match an empty string.
349        (^a|^)+   was not marked as anchored.
350        (.*a|.*)+ was not marked as matching at start or after a newline.
351    
352    12. Yet more internal refactoring has removed another argument from the match()
353        function. Special calls to this function are now indicated by setting a
354        value in a variable in the "match data" data block.
355    
356    13. Be more explicit in pcre_study() instead of relying on "default" for
357        opcodes that mean there is no starting character; this means that when new
358        ones are added and accidentally left out of pcre_study(), testing should
359        pick them up.
360    
361    14. The -s option of pcretest has been documented for ages as being an old
362        synonym of -m (show memory usage). I have changed it to mean "force study
363        for every regex", that is, assume /S for every regex. This is similar to -i
364        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
365        using it. It makes it easier to run collections of tests with and without
366        study enabled, and thereby test pcre_study() more easily. All the standard
367        tests are now run with and without -s (but some patterns can be marked as
368        "never study" - see 20 below).
369    
370    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
371        restoration of the capturing data to the outer values was not happening
372        correctly.
373    
374    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
375        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
376        pattern had matched an empty string, and so incorrectly returned a no
377        match.
378    
379    17. There was optimizing code for the last branch of non-capturing parentheses,
380        and also for the obeyed branch of a conditional subexpression, which used
381        tail recursion to cut down on stack usage. Unfortunately, now that there is
382        the possibility of (*THEN) occurring in these branches, tail recursion is
383        no longer possible because the return has to be checked for (*THEN). These
384        two optimizations have therefore been removed. [But see 8.20/11 above.]
385    
386    18. If a pattern containing \R was studied, it was assumed that \R always
387        matched two bytes, thus causing the minimum subject length to be
388        incorrectly computed because \R can also match just one byte.
389    
390    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
391        was incorrectly computed.
392    
393    20. If /S is present twice on a test pattern in pcretest input, it now
394        *disables* studying, thereby overriding the use of -s on the command line
395        (see 14 above). This is necessary for one or two tests to keep the output
396        identical in both cases.
397    
398    21. When (*ACCEPT) was used in an assertion that matched an empty string and
399        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
400    
401    22. When an atomic group that contained a capturing parenthesis was
402        successfully matched, but the branch in which it appeared failed, the
403        capturing was not being forgotten if a higher numbered group was later
404        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
405        group 1 to "a", when in fact it should be unset. This applied to multi-
406        branched capturing and non-capturing groups, repeated or not, and also to
407        positive assertions (capturing in negative assertions does not happen
408        in PCRE) and also to nested atomic groups.
409    
410    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
411        subject after a captured substring, to make it easier to tell which of a
412        number of identical substrings has been captured.
413    
414    24. The way atomic groups are processed by pcre_exec() has been changed so that
415        if they are repeated, backtracking one repetition now resets captured
416        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
417        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
418        "aaa". Previously, it would have been "a". As part of this code
419        refactoring, the way recursive calls are handled has also been changed.
420    
421    25. If an assertion condition captured any substrings, they were not passed
422        back unless some other capturing happened later. For example, if
423        (?(?=(a))a) was matched against "a", no capturing was returned.
424    
425    26. When studying a pattern that contained subroutine calls or assertions,
426        the code for finding the minimum length of a possible match was handling
427        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
428        group 1 called group 2 while simultaneously a separate group 2 called group
429        1). A stack overflow occurred in this case. I have fixed this by limiting
430        the recursion depth to 10.
431    
432    27. Updated RunTest.bat in the distribution to the version supplied by Tom
433        Fortmann. This supports explicit test numbers on the command line, and has
434        argument validation and error reporting.
435    
436    28. An instance of \X with an unlimited repeat could fail if at any point the
437        first character it looked at was a mark character.
438    
439    29. Some minor code refactoring concerning Unicode properties and scripts
440        should reduce the stack requirement of match() slightly.
441    
442    30. Added the '=' option to pcretest to check the setting of unused capturing
443        slots at the end of the pattern, which are documented as being -1, but are
444        not included in the return count.
445    
446    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
447        compiled something random. Now it gives a compile-time error (as does
448        Perl).
449    
450    32. A *MARK encountered during the processing of a positive assertion is now
451        recorded and passed back (compatible with Perl).
452    
453    33. If --only-matching or --colour was set on a pcregrep call whose pattern
454        had alternative anchored branches, the search for a second match in a line
455        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
456        matched the line "0102" twice. The same bug affected patterns that started
457        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
458        twice.
459    
460    34. Previously, PCRE did not allow quantification of assertions. However, Perl
461        does, and because of capturing effects, quantifying parenthesized
462        assertions may at times be useful. Quantifiers are now allowed for
463        parenthesized assertions.
464    
465    35. A minor code tidy in pcre_compile() when checking options for \R usage.
466    
467    36. \g was being checked for fancy things in a character class, when it should
468        just be a literal "g".
469    
470    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
471        appearance of a nested POSIX class supersedes an apparent external class.
472        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
473        unescaped square brackets may also appear as part of class names. For
474        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
475        more like Perl. (But see 8.20/1 above.)
476    
477    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
478        was because it thought it was \N{name}, which is not supported).
479    
480    39. Add minix to OS list not supporting the -S option in pcretest.
481    
482    40. PCRE tries to detect cases of infinite recursion at compile time, but it
483        cannot analyze patterns in sufficient detail to catch mutual recursions
484        such as ((?1))((?2)). There is now a runtime test that gives an error if a
485        subgroup is called recursively as a subpattern for a second time at the
486        same position in the subject string. In previous releases this might have
487        been caught by the recursion limit, or it might have run out of stack.
488    
489    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
490        happen only once. PCRE was, however incorrectly giving a compile time error
491        "recursive call could loop indefinitely" because it cannot analyze the
492        pattern in sufficient detail. The compile time test no longer happens when
493        PCRE is compiling a conditional subpattern, but actual runaway loops are
494        now caught at runtime (see 40 above).
495    
496    42. It seems that Perl allows any characters other than a closing parenthesis
497        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
498        has been changed to be the same.
499    
500    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
501        as not to get warnings when autogen.sh is called. Also changed
502        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
503    
504    44. To help people who use pcregrep to scan files containing exceedingly long
505        lines, the following changes have been made:
506    
507        (a) The default value of the buffer size parameter has been increased from
508            8K to 20K. (The actual buffer used is three times this size.)
509    
510        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
511            PCRE is built.
512    
513        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
514            to be set at run time.
515    
516        (d) Numerical values in pcregrep options can be followed by K or M, for
517            example --buffer-size=50K.
518    
519        (e) If a line being scanned overflows pcregrep's buffer, an error is now
520            given and the return code is set to 2.
521    
522    45. Add a pointer to the latest mark to the callout data block.
523    
524    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
525        partial match of an empty string instead of no match. This was specific to
526        the use of ".".
527    
528    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
529        complete match instead of a partial match. This bug was dependent on both
530        the PCRE_UTF8 and PCRE_DOTALL options being set.
531    
532    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
533        starting byte set, because \b was not being ignored.
534    
535    
536    Version 8.12 15-Jan-2011
537    ------------------------
538    
539    1.  Fixed some typos in the markup of the man pages, and wrote a script that
540        checks for such things as part of the documentation building process.
541    
542    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
543        --match-limit and --recursion-limit options (added for 8.11). In
544        particular, this made one of the standard tests fail. (The integer value
545        went into the wrong half of a long int.)
546    
547    3.  If the --colour option was given to pcregrep with -v (invert match), it
548        did strange things, either producing crazy output, or crashing. It should,
549        of course, ignore a request for colour when reporting lines that do not
550        match.
551    
552    4.  Another pcregrep bug caused similar problems if --colour was specified with
553        -M (multiline) and the pattern match finished with a line ending.
554    
555    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
556        matched in multiline mode, the following line was shown as part of the
557        match. This seems wrong, so I have changed it.
558    
559    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
560        the check for further matches in the same line (so they could be coloured)
561        to overrun the end of the current line. If another match was found, it was
562        incorrectly shown (and then shown again when found in the next line).
563    
564    7.  If pcregrep was compiled under Windows, there was a reference to the
565        function pcregrep_exit() before it was defined. I am assuming this was
566        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
567        reported by a user. I've moved the definition above the reference.
568    
569    
570    Version 8.11 10-Dec-2010
571    ------------------------
572    
573    1.  (*THEN) was not working properly if there were untried alternatives prior
574        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
575        backtracked to try for "b" instead of moving to the next alternative branch
576        at the same level (in this case, to look for "c"). The Perl documentation
577        is clear that when (*THEN) is backtracked onto, it goes to the "next
578        alternative in the innermost enclosing group".
579    
580    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
581        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
582        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
583        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
584        (*THEN).
585    
586    3.  If \s appeared in a character class, it removed the VT character from
587        the class, even if it had been included by some previous item, for example
588        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
589        of \s, but is part of the POSIX "space" class.)
590    
591    4.  A partial match never returns an empty string (because you can always
592        match an empty string at the end of the subject); however the checking for
593        an empty string was starting at the "start of match" point. This has been
594        changed to the "earliest inspected character" point, because the returned
595        data for a partial match starts at this character. This means that, for
596        example, /(?<=abc)def/ gives a partial match for the subject "abc"
597        (previously it gave "no match").
598    
599    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
600        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
601        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
602        has an implication that the given string is incomplete (because a partial
603        match is preferred over a full match). For this reason, these items now
604        give a partial match in this situation. [Aside: previously, the one case
605        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
606        match rather than a full match, which was wrong by the old rules, but is
607        now correct.]
608    
609    6.  There was a bug in the handling of #-introduced comments, recognized when
610        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
611        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
612        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
613        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
614        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
615        places in pcre_compile().
616    
617    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
618        comments when looking ahead for named forward references to subpatterns,
619        the only newline sequence it recognized was NL. It now handles newlines
620        according to the set newline convention.
621    
622    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
623        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
624        cater for a lack of strerror(). These oversights have been fixed.
625    
626    9.  Added --match-limit and --recursion-limit to pcregrep.
627    
628    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
629    
630    11. When the -o option was used, pcregrep was setting a return code of 1, even
631        when matches were found, and --line-buffered was not being honoured.
632    
633    12. Added an optional parentheses number to the -o and --only-matching options
634        of pcregrep.
635    
636    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
637        can match an empty string. The code to do it in pcretest and pcredemo
638        needed fixing:
639    
640        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
641            only one byte after an empty string match just before CRLF (this case
642            just got forgotten; "any" and "anycrlf" were OK).
643    
644        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
645            mode when an empty string match preceded an ASCII character followed by
646            a non-ASCII character. (The code for advancing by one character rather
647            than one byte was nonsense.)
648    
649        (c) The pcredemo.c sample program did not have any code at all to handle
650            the cases when CRLF is a valid newline sequence.
651    
652    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
653        as a starting offset was within the subject string. There is now a new
654        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
655        negative or greater than the length of the string. In order to test this,
656        pcretest is extended to allow the setting of negative starting offsets.
657    
658    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
659        starting offset points to the beginning of a UTF-8 character was
660        unnecessarily clumsy. I tidied it up.
661    
662    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
663        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
664    
665    17. Nobody had reported that the --include_dir option, which was added in
666        release 7.7 should have been called --include-dir (hyphen, not underscore)
667        for compatibility with GNU grep. I have changed it to --include-dir, but
668        left --include_dir as an undocumented synonym, and the same for
669        --exclude-dir, though that is not available in GNU grep, at least as of
670        release 2.5.4.
671    
672    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
673        characters from a string of bytes) have been redefined so as not to use
674        loops, in order to improve performance in some environments. At the same
675        time, I abstracted some of the common code into auxiliary macros to save
676        repetition (this should not affect the compiled code).
677    
678    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
679        compile-time error is now given if \c is not followed by an ASCII
680        character, that is, a byte less than 128. (In EBCDIC mode, the code is
681        different, and any byte value is allowed.)
682    
683    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
684        START_OPTIMIZE option, which is now allowed at compile time - but just
685        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
686        to pcregrep and other applications that have no direct access to PCRE
687        options. The new /Y option in pcretest sets this option when calling
688        pcre_compile().
689    
690    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
691        back references. Groups containing recursive back references were forced to
692        be atomic by that change, but in the case of named groups, the amount of
693        memory required was incorrectly computed, leading to "Failed: internal
694        error: code overflow". This has been fixed.
695    
696    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
697        pcretest.c, to avoid build problems in some Borland environments.
698    
699    
700    Version 8.10 25-Jun-2010
701    ------------------------
702    
703    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
704        THEN.
705    
706    2.  (*ACCEPT) was not working when inside an atomic group.
707    
708    3.  Inside a character class, \B is treated as a literal by default, but
709        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
710        causes the error). The code is unchanged, but I tidied the documentation.
711    
712    4.  Inside a character class, PCRE always treated \R and \X as literals,
713        whereas Perl faults them if its -w option is set. I have changed PCRE so
714        that it faults them when PCRE_EXTRA is set.
715    
716    5.  Added support for \N, which always matches any character other than
717        newline. (It is the same as "." when PCRE_DOTALL is not set.)
718    
719    6.  When compiling pcregrep with newer versions of gcc which may have
720        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
721        declared with attribute warn_unused_result" were given. Just casting the
722        result to (void) does not stop the warnings; a more elaborate fudge is
723        needed. I've used a macro to implement this.
724    
725    7.  Minor change to pcretest.c to avoid a compiler warning.
726    
727    8.  Added four artifical Unicode properties to help with an option to make
728        \s etc use properties (see next item). The new properties are: Xan
729        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
730    
731    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
732        use Unicode properties. (*UCP) at the start of a pattern can be used to set
733        this option. Modified pcretest to add /W to test this facility. Added
734        REG_UCP to make it available via the POSIX interface.
735    
736    10. Added --line-buffered to pcregrep.
737    
738    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
739        studied, and the match started with a letter with a code point greater than
740        127 whose first byte was different to the first byte of the other case of
741        the letter, the other case of this starting letter was not recognized
742        (#976).
743    
744    12. If a pattern that was studied started with a repeated Unicode property
745        test, for example, \p{Nd}+, there was the theoretical possibility of
746        setting up an incorrect bitmap of starting bytes, but fortunately it could
747        not have actually happened in practice until change 8 above was made (it
748        added property types that matched character-matching opcodes).
749    
750    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
751        possible starting bytes for non-anchored patterns.
752    
753    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
754        \R, and also a number of cases that involve Unicode properties, both
755        explicit and implicit when PCRE_UCP is set.
756    
757    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
758        input, it could crash or give wrong results if characters with values
759        greater than 0xc0 were present in the subject string. (Detail: it assumed
760        UTF-8 input when processing these items.)
761    
762    16. Added a lot of (int) casts to avoid compiler warnings in systems where
763        size_t is 64-bit (#991).
764    
765    17. Added a check for running out of memory when PCRE is compiled with
766        --disable-stack-for-recursion (#990).
767    
768    18. If the last data line in a file for pcretest does not have a newline on
769        the end, a newline was missing in the output.
770    
771    19. The default pcre_chartables.c file recognizes only ASCII characters (values
772        less than 128) in its various bitmaps. However, there is a facility for
773        generating tables according to the current locale when PCRE is compiled. It
774        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
775        characters, are recognized by isspace() and therefore were getting set in
776        these tables, and indeed these tables seem to approximate to ISO 8859. This
777        caused a problem in UTF-8 mode when pcre_study() was used to create a list
778        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
779        which of course cannot start UTF-8 characters. I have changed the code so
780        that only real ASCII characters (less than 128) and the correct starting
781        bytes for UTF-8 encodings are set for characters greater than 127 when in
782        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
783        altogether.)
784    
785    20. Added the /T option to pcretest so as to be able to run tests with non-
786        standard character tables, thus making it possible to include the tests
787        used for 19 above in the standard set of tests.
788    
789    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
790        reference to a subpattern the other side of a comment that contains an
791        opening parenthesis caused either an internal compiling error, or a
792        reference to the wrong subpattern.
793    
794    
795    Version 8.02 19-Mar-2010
796    ------------------------
797    
798    1.  The Unicode data tables have been updated to Unicode 5.2.0.
799    
800    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
801        configured.
802    
803    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
804        original author of that file, following a query about its status.
805    
806    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
807        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
808    
809    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
810        quantifier applied to a forward-referencing subroutine call, could compile
811        incorrect code or give the error "internal error: previously-checked
812        referenced subpattern not found".
813    
814    6.  Both MS Visual Studio and Symbian OS have problems with initializing
815        variables to point to external functions. For these systems, therefore,
816        pcre_malloc etc. are now initialized to local functions that call the
817        relevant global functions.
818    
819    7.  There were two entries missing in the vectors called coptable and poptable
820        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
821        I've fixed the data, and added a kludgy way of testing at compile time that
822        the lengths are correct (equal to the number of opcodes).
823    
824    8.  Following on from 7, I added a similar kludge to check the length of the
825        eint vector in pcreposix.c.
826    
827    9.  Error texts for pcre_compile() are held as one long string to avoid too
828        much relocation at load time. To find a text, the string is searched,
829        counting zeros. There was no check for running off the end of the string,
830        which could happen if a new error number was added without updating the
831        string.
832    
833    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
834    
835    11. \K was not working if it appeared in an atomic group or in a group that
836        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
837        \K is "not well defined" if used in an assertion. PCRE now accepts it if
838        the assertion is positive, but not if it is negative.
839    
840    12. Change 11 fortuitously reduced the size of the stack frame used in the
841        "match()" function of pcre_exec.c by one pointer. Forthcoming
842        implementation of support for (*MARK) will need an extra pointer on the
843        stack; I have reserved it now, so that the stack frame size does not
844        decrease.
845    
846    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
847        item in branch that calls a recursion is a subroutine call - as in the
848        second branch in the above example - was incorrectly given the compile-
849        time error "recursive call could loop indefinitely" because pcre_compile()
850        was not correctly checking the subroutine for matching a non-empty string.
851    
852    14. The checks for overrunning compiling workspace could trigger after an
853        overrun had occurred. This is a "should never occur" error, but it can be
854        triggered by pathological patterns such as hundreds of nested parentheses.
855        The checks now trigger 100 bytes before the end of the workspace.
856    
857    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
858    
859    
860    Version 8.01 19-Jan-2010
861    ------------------------
862    
863    1.  If a pattern contained a conditional subpattern with only one branch (in
864        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
865        computed the wrong minimum data length (which is of course zero for such
866        subpatterns). This could cause incorrect "no match" results.
867    
868    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
869        the pattern is reset in the first branch, pcre_compile() failed with
870        "internal error: code overflow at offset...". This happened only when
871        the reset was to the original external option setting. (An optimization
872        abstracts leading options settings into an external setting, which was the
873        cause of this.)
874    
875    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
876        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
877        assertion pattern did not match (meaning that the assertion was true), it
878        was incorrectly treated as false if the SKIP had been reached during the
879        matching. This also applied to assertions used as conditions.
880    
881    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
882        assertion subpattern, including such a pattern used as a condition,
883        unpredictable results occurred, instead of the error return
884        PCRE_ERROR_DFA_UITEM.
885    
886    5.  The C++ GlobalReplace function was not working like Perl for the special
887        situation when an empty string is matched. It now does the fancy magic
888        stuff that is necessary.
889    
890    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
891        removed. (These were left over from very, very early versions of PCRE.)
892    
893    7.  Some cosmetic changes to the code to make life easier when compiling it
894        as part of something else:
895    
896        (a) Change DEBUG to PCRE_DEBUG.
897    
898        (b) In pcre_compile(), rename the member of the "branch_chain" structure
899            called "current" as "current_branch", to prevent a collision with the
900            Linux macro when compiled as a kernel module.
901    
902        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
903            prevent a collision with the Linux macro when compiled as a kernel
904            module.
905    
906    8.  In pcre_compile() there are some checks for integer overflows that used to
907        cast potentially large values to (double). This has been changed to that
908        when building, a check for int64_t is made, and if it is found, it is used
909        instead, thus avoiding the use of floating point arithmetic. (There is no
910        other use of FP in PCRE.) If int64_t is not found, the fallback is to
911        double.
912    
913    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
914        2005 (difference between two addresses compared to an unsigned value).
915    
916    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
917        custom one, because of the following reported problem in Windows:
918    
919          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
920              under Win32.
921          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
922              therefore missing the function definition.
923          - The compiler thus generates a "C" signature for the test function.
924          - The linker fails to find the "C" function.
925          - PCRE fails to configure if asked to do so against libbz2.
926    
927    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
928        messages were output:
929    
930          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
931          rerunning libtoolize, to keep the correct libtool macros in-tree.
932          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
933    
934        I have done both of these things.
935    
936    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
937        most of the time, it *can* run out if it is given a pattern that contains a
938        runaway infinite recursion. I updated the discussion in the pcrestack man
939        page.
940    
941    13. Now that we have gone to the x.xx style of version numbers, the minor
942        version may start with zero. Using 08 or 09 is a bad idea because users
943        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
944        interpreted as invalid octal numbers. I've updated the previous comment in
945        configure.ac, and also added a check that gives an error if 08 or 09 are
946        used.
947    
948    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
949        causing partial matching to fail when the end of the subject matched \W
950        in a UTF-8 pattern where \W was quantified with a minimum of 3.
951    
952    15. There were some discrepancies between the declarations in pcre_internal.h
953        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
954        their definitions. The declarations used "const uschar *" and the
955        definitions used USPTR. Even though USPTR is normally defined as "const
956        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
957        reported that: "This difference in casting confuses some C++ compilers, for
958        example, SunCC recognizes above declarations as different functions and
959        generates broken code for hbpcre." I have changed the declarations to use
960        USPTR.
961    
962    16. GNU libtool is named differently on some systems. The autogen.sh script now
963        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
964        (FreeBSD).
965    
966    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
967        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
968        comment: "Figure out how to create a longlong from a string: strtoll and
969        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
970        instance, but it only takes 2 args instead of 3!"
971    
972    18. A subtle bug concerned with back references has been fixed by a change of
973        specification, with a corresponding code fix. A pattern such as
974        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
975        refers, was giving matches when it shouldn't. For example, xa=xaaa would
976        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
977        same bug. Such groups have to be quantified to be useful, or contained
978        inside another quantified group. (If there's no repetition, the reference
979        can never match.) The problem arises because, having left the group and
980        moved on to the rest of the pattern, a later failure that backtracks into
981        the group uses the captured value from the final iteration of the group
982        rather than the correct earlier one. I have fixed this in PCRE by forcing
983        any group that contains a reference to itself to be an atomic group; that
984        is, there cannot be any backtracking into it once it has completed. This is
985        similar to recursive and subroutine calls.
986    
987    
988    Version 8.00 19-Oct-09
989    ----------------------
990    
991    1.  The table for translating pcre_compile() error codes into POSIX error codes
992        was out-of-date, and there was no check on the pcre_compile() error code
993        being within the table. This could lead to an OK return being given in
994        error.
995    
996    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
997        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
998        in a Windows environment.
999    
1000    3.  The pcregrep --count option prints the count for each file even when it is
1001        zero, as does GNU grep. However, pcregrep was also printing all files when
1002        --files-with-matches was added. Now, when both options are given, it prints
1003        counts only for those files that have at least one match. (GNU grep just
1004        prints the file name in this circumstance, but including the count seems
1005        more useful - otherwise, why use --count?) Also ensured that the
1006        combination -clh just lists non-zero counts, with no names.
1007    
1008    4.  The long form of the pcregrep -F option was incorrectly implemented as
1009        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1010        but it seems right to fix it, and I didn't think it was worth preserving
1011        the old behaviour.
1012    
1013    5.  The command line items --regex=pattern and --regexp=pattern were not
1014        recognized by pcregrep, which required --regex pattern or --regexp pattern
1015        (with a space rather than an '='). The man page documented the '=' forms,
1016        which are compatible with GNU grep; these now work.
1017    
1018    6.  No libpcreposix.pc file was created for pkg-config; there was just
1019        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1020    
1021    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1022        when UCP support is not needed, by modifying the Python script that
1023        generates it from Unicode data files. This should not matter if the module
1024        is correctly used as a library, but I received one complaint about 50K of
1025        unwanted data. My guess is that the person linked everything into his
1026        program rather than using a library. Anyway, it does no harm.
1027    
1028    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1029        was a minimum greater than 1 for a wide character in a possessive
1030        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1031        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1032        character. Chaos in the form of incorrect output or a compiling loop could
1033        result.
1034    
1035    9.  The restrictions on what a pattern can contain when partial matching is
1036        requested for pcre_exec() have been removed. All patterns can now be
1037        partially matched by this function. In addition, if there are at least two
1038        slots in the offset vector, the offset of the earliest inspected character
1039        for the match and the offset of the end of the subject are set in them when
1040        PCRE_ERROR_PARTIAL is returned.
1041    
1042    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1043        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1044        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1045        and may be more useful for multi-segment matching.
1046    
1047    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1048        used to be given if ever the end of the subject was reached; now it is
1049        given only if matching could not proceed because another character was
1050        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1051        string "Z", which now yields "no match" instead of "partial match". In the
1052        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1053        final character ended with (*FAIL).
1054    
1055    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1056        if the pattern had a "must contain" character that was already found in the
1057        earlier partial match, unless partial matching was again requested. For
1058        example, with the pattern /dog.(body)?/, the "must contain" character is
1059        "g". If the first part-match was for the string "dog", restarting with
1060        "sbody" failed. This bug has been fixed.
1061    
1062    13. The string returned by pcre_dfa_exec() after a partial match has been
1063        changed so that it starts at the first inspected character rather than the
1064        first character of the match. This makes a difference only if the pattern
1065        starts with a lookbehind assertion or \b or \B (\K is not supported by
1066        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1067        matching functions compatible, and I think it's the right thing to do.
1068    
1069    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1070        so that the demonstration program is easily available in environments where
1071        PCRE has not been installed from source.
1072    
1073    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1074        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1075        library.
1076    
1077    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1078        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1079        is not the first non-POSIX option to be added. Clearly some people find
1080        these options useful.
1081    
1082    17. If a caller to the POSIX matching function regexec() passes a non-zero
1083        value for nmatch with a NULL value for pmatch, the value of
1084        nmatch is forced to zero.
1085    
1086    18. RunGrepTest did not have a test for the availability of the -u option of
1087        the diff command, as RunTest does. It now checks in the same way as
1088        RunTest, and also checks for the -b option.
1089    
1090    19. If an odd number of negated classes containing just a single character
1091        interposed, within parentheses, between a forward reference to a named
1092        subpattern and the definition of the subpattern, compilation crashed with
1093        an internal error, complaining that it could not find the referenced
1094        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1095        [The bug was that it was starting one character too far in when skipping
1096        over the character class, thus treating the ] as data rather than
1097        terminating the class. This meant it could skip too much.]
1098    
1099    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1100        /g option in pcretest when the pattern contains \K, which makes it possible
1101        to have an empty string match not at the start, even when the pattern is
1102        anchored. Updated pcretest and pcredemo to use this option.
1103    
1104    21. If the maximum number of capturing subpatterns in a recursion was greater
1105        than the maximum at the outer level, the higher number was returned, but
1106        with unset values at the outer level. The correct (outer level) value is
1107        now given.
1108    
1109    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1110        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1111        make it do so. The string so far is captured, making this feature
1112        compatible with Perl.
1113    
1114    23. The tests have been re-organized, adding tests 11 and 12, to make it
1115        possible to check the Perl 5.10 features against Perl 5.10.
1116    
1117    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1118        pattern matches a fixed length string. PCRE did not allow this; now it
1119        does. Neither allows recursion.
1120    
1121    25. I finally figured out how to implement a request to provide the minimum
1122        length of subject string that was needed in order to match a given pattern.
1123        (It was back references and recursion that I had previously got hung up
1124        on.) This code has now been added to pcre_study(); it finds a lower bound
1125        to the length of subject needed. It is not necessarily the greatest lower
1126        bound, but using it to avoid searching strings that are too short does give
1127        some useful speed-ups. The value is available to calling programs via
1128        pcre_fullinfo().
1129    
1130    26. While implementing 25, I discovered to my embarrassment that pcretest had
1131        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1132        study optimizations had never been tested with that matching function.
1133        Oops. What is worse, even when it was passed study data, there was a bug in
1134        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1135        were also very few tests of studied patterns with pcre_dfa_exec().
1136    
1137    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1138        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1139        on the other side of the coin, they are no longer allowed to have different
1140        names, because these cannot be distinguished in PCRE, and this has caused
1141        confusion. (This is a difference from Perl.)
1142    
1143    28. When duplicate subpattern names are present (necessarily with different
1144        numbers, as required by 27 above), and a test is made by name in a
1145        conditional pattern, either for a subpattern having been matched, or for
1146        recursion in such a pattern, all the associated numbered subpatterns are
1147        tested, and the overall condition is true if the condition is true for any
1148        one of them. This is the way Perl works, and is also more like the way
1149        testing by number works.
1150    
1151    
1152    Version 7.9 11-Apr-09
1153    ---------------------
1154    
1155    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1156        (pcretest), all targets were linked against these libraries. This included
1157        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1158        libraries. This caused unwanted dependencies to be created. This problem
1159        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1160        pcretest is linked with readline.
1161    
1162    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1163        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1164        moved outside it again, because FALSE and TRUE are already defined in AIX,
1165        but BOOL is not.
1166    
1167    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1168        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1169    
1170    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1171        hyphens) following file names and line numbers when outputting matching
1172        lines. This is not true; no spaces are inserted. I have also clarified the
1173        wording for the --colour (or --color) option.
1174    
1175    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1176        was not coloured; this is different to GNU grep, so I have changed it to be
1177        the same.
1178    
1179    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1180        each matching line was coloured. Now it goes on to look for further matches
1181        of any of the test patterns, which is the same behaviour as GNU grep.
1182    
1183    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1184        doesn't make sense to accept an empty string match in pcregrep, so I have
1185        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1186        seems to be how GNU grep behaves.
1187    
1188    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1189        start or after a newline", because the conditional assertion was not being
1190        correctly handled. The rule now is that both the assertion and what follows
1191        in the first alternative must satisfy the test.
1192    
1193    9.  If auto-callout was enabled in a pattern with a conditional group whose
1194        condition was an assertion, PCRE could crash during matching, both with
1195        pcre_exec() and pcre_dfa_exec().
1196    
1197    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1198        used for matching.
1199    
1200    11. Unicode property support in character classes was not working for
1201        characters (bytes) greater than 127 when not in UTF-8 mode.
1202    
1203    12. Added the -M command line option to pcretest.
1204    
1205    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1206    
1207    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1208    
1209    16. Added comments and documentation about mis-use of no_arg in the C++
1210        wrapper.
1211    
1212    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1213        from Martin Jerabek that uses macro names for all relevant character and
1214        string constants.
1215    
1216    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1217        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1218        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1219        these, but not everybody uses configure.
1220    
1221    19. A conditional group that had only one branch was not being correctly
1222        recognized as an item that could match an empty string. This meant that an
1223        enclosing group might also not be so recognized, causing infinite looping
1224        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1225        with the subject "ab", where knowledge that the repeated group can match
1226        nothing is needed in order to break the loop.
1227    
1228    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1229        exec(), but without supplying a callout function, matching went wrong.
1230    
1231    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1232        leak if the size of the offset vector was greater than 30. When the vector
1233        is smaller, the saved offsets during recursion go onto a local stack
1234        vector, but for larger vectors malloc() is used. It was failing to free
1235        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1236        error, in fact).
1237    
1238    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1239        heapframe that is used only when UTF-8 support is enabled. This caused no
1240        problem, but was untidy.
1241    
1242    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1243        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1244        included within another project.
1245    
1246    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1247        slightly modified by me:
1248    
1249          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1250              not building pcregrep.
1251    
1252          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1253              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1254    
1255    25. Forward references, both numeric and by name, in patterns that made use of
1256        duplicate group numbers, could behave incorrectly or give incorrect errors,
1257        because when scanning forward to find the reference group, PCRE was not
1258        taking into account the duplicate group numbers. A pattern such as
1259        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1260    
1261    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1262        the feature of a custom pointer more persuasive (as requested by a user).
1263    
1264    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1265        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1266        pre-defined.
1267    
1268    28. Added support for (*UTF8) at the start of a pattern.
1269    
1270    29. Arrange for flags added by the "release type" setting in CMake to be shown
1271        in the configuration summary.
1272    
1273    
1274    Version 7.8 05-Sep-08
1275    ---------------------
1276    
1277    1.  Replaced UCP searching code with optimized version as implemented for Ad
1278        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1279        stage table and inline lookup instead of a function, giving speed ups of 2
1280        to 5 times on some simple patterns that I tested. Permission was given to
1281        distribute the MultiStage2.py script that generates the tables (it's not in
1282        the tarball, but is in the Subversion repository).
1283    
1284    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1285        scripts.
1286    
1287    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1288        a group with a zero qualifier. The result of the study could be incorrect,
1289        or the function might crash, depending on the pattern.
1290    
1291    4.  Caseless matching was not working for non-ASCII characters in back
1292        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1293        It now works when Unicode Property Support is available.
1294    
1295    5.  In pcretest, an escape such as \x{de} in the data was always generating
1296        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1297        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1298        truncation.
1299    
1300    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1301    
1302    7.  Added two (int) casts to pcregrep when printing the difference of two
1303        pointers, in case they are 64-bit values.
1304    
1305    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1306        test 2 if it fails.
1307    
1308    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1309        and a #define of that name to empty if it is not externally set. This is to
1310        allow users of MSVC to set it if necessary.
1311    
1312    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1313        the convenience functions in the pcre_get.c source file.
1314    
1315    11. An option change at the start of a pattern that had top-level alternatives
1316        could cause overwriting and/or a crash. This command provoked a crash in
1317        some environments:
1318    
1319          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1320    
1321        This potential security problem was recorded as CVE-2008-2371.
1322    
1323    12. For a pattern where the match had to start at the beginning or immediately
1324        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1325        pcre_dfa_exec() could read past the end of the passed subject if there was
1326        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1327        pcretest so that it places the subject at the end of its malloc-ed buffer.
1328    
1329    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1330        exec() might read past the end of the data buffer in UTF-8 mode.
1331    
1332    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1333        the data contained the byte 0x85 as part of a UTF-8 character within its
1334        first line. This applied both to normal and DFA matching.
1335    
1336    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1337        /^[^d]*?$/8 failed to match "abc".
1338    
1339    16. Added a missing copyright notice to pcrecpp_internal.h.
1340    
1341    17. Make it more clear in the documentation that values returned from
1342        pcre_exec() in ovector are byte offsets, not character counts.
1343    
1344    18. Tidied a few places to stop certain compilers from issuing warnings.
1345    
1346    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1347        supplied by Stefan Weber. I made a further small update for 7.8 because
1348        there is a change of source arrangements: the pcre_searchfuncs.c module is
1349        replaced by pcre_ucd.c.
1350    
1351    
1352    Version 7.7 07-May-08
1353    ---------------------
1354    
1355    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1356        a string to a long long, pretend we don't even have a long long." This is
1357        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1358    
1359    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1360        pre-7.6 versions, which defined a global no_arg variable instead of putting
1361        it in the RE class. (See also #8 below.)
1362    
1363    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1364        Lopes.
1365    
1366    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1367    
1368        (1) The include/exclude patterns were being applied to the whole pathnames
1369            of files, instead of just to the final components.
1370    
1371        (2) If there was more than one level of directory, the subdirectories were
1372            skipped unless they satisfied the include/exclude conditions. This is
1373            inconsistent with GNU grep (and could even be seen as contrary to the
1374            pcregrep specification - which I improved to make it absolutely clear).
1375            The action now is always to scan all levels of directory, and just
1376            apply the include/exclude patterns to regular files.
1377    
1378    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1379        --exclude_dir in the tests to avoid scanning .svn directories.
1380    
1381    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1382        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1383        doesn't support NULs in patterns.
1384    
1385    7.  Added some missing "const"s to declarations of static tables in
1386        pcre_compile.c and pcre_dfa_exec.c.
1387    
1388    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1389        caused by fix #2  above. (Subsequently also a second patch to fix the
1390        first patch. And a third patch - this was a messy problem.)
1391    
1392    9.  Applied Craig's patch to remove the use of push_back().
1393    
1394    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1395        matching function regexec().
1396    
1397    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1398        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1399        references. PCRE supports relative numbers with this syntax (I don't think
1400        Oniguruma does).
1401    
1402    12. Previously, a group with a zero repeat such as (...){0} was completely
1403        omitted from the compiled regex. However, this means that if the group
1404        was called as a subroutine from elsewhere in the pattern, things went wrong
1405        (an internal error was given). Such groups are now left in the compiled
1406        pattern, with a new opcode that causes them to be skipped at execution
1407        time.
1408    
1409    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1410        to the way PCRE behaves:
1411    
1412        (a) A lone ] character is dis-allowed (Perl treats it as data).
1413    
1414        (b) A back reference to an unmatched subpattern matches an empty string
1415            (Perl fails the current match path).
1416    
1417        (c) A data ] in a character class must be notated as \] because if the
1418            first data character in a class is ], it defines an empty class. (In
1419            Perl it is not possible to have an empty class.) The empty class []
1420            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1421            The negative empty class [^] matches any one character, independently
1422            of the DOTALL setting.
1423    
1424    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1425        non-existent subpattern following a character class starting with ']' and
1426        containing () gave an internal compiling error instead of "reference to
1427        non-existent subpattern". Fortunately, when the pattern did exist, the
1428        compiled code was correct. (When scanning forwards to check for the
1429        existencd of the subpattern, it was treating the data ']' as terminating
1430        the class, so got the count wrong. When actually compiling, the reference
1431        was subsequently set up correctly.)
1432    
1433    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1434        it was being rejected as not supported by pcre_dfa_exec(), even though
1435        other assertions are supported. I have made pcre_dfa_exec() support
1436        (*FAIL).
1437    
1438    16. The implementation of 13c above involved the invention of a new opcode,
1439        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1440        cannot be changed at match time, I realized I could make a small
1441        improvement to matching performance by compiling OP_ALLANY instead of
1442        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1443        on the OP_ANY path.
1444    
1445    17. Compiling pcretest on Windows with readline support failed without the
1446        following two fixes: (1) Make the unistd.h include conditional on
1447        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1448    
1449    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1450        ncurses library to be included for pcretest when ReadLine support is
1451        requested, but also to allow for it to be overridden. This patch came from
1452        Daniel Bergström.
1453    
1454    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1455        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1456        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1457        spotting this.
1458    
1459    
1460    Version 7.6 28-Jan-08
1461    ---------------------
1462    
1463    1.  A character class containing a very large number of characters with
1464        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1465        overflow.
1466    
1467    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1468        HAVE_LONG_LONG is not defined.
1469    
1470    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1471        bring them up to date and include new features. This patch includes:
1472    
1473        - Fixed PH's badly added libz and libbz2 support.
1474        - Fixed a problem with static linking.
1475        - Added pcredemo. [But later removed - see 7 below.]
1476        - Fixed dftables problem and added an option.
1477        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1478            HAVE_LONG_LONG.
1479        - Added readline support for pcretest.
1480        - Added an listing of the option settings after cmake has run.
1481    
1482    4.  A user submitted a patch to Makefile that makes it easy to create
1483        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1484        Makefile.am that cause it to include this special target, without
1485        affecting anything else. Note that the same mingw target plus all
1486        the other distribution libraries and programs are now supported
1487        when configuring with CMake (see 6 below) instead of with
1488        Configure/Make.
1489    
1490    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1491        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1492        exported in the Windows port". It has not yet been confirmed that the patch
1493        solves the problem, but it does no harm.
1494    
1495    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1496        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1497        with CMake, and also correct the comment about stack recursion.
1498    
1499    7.  Remove the automatic building of pcredemo from the ./configure system and
1500        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1501        of a program that users should build themselves after PCRE is installed, so
1502        building it automatically is not really right. What is more, it gave
1503        trouble in some build environments.
1504    
1505    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1506    
1507    
1508    Version 7.5 10-Jan-08
1509    ---------------------
1510    
1511    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1512        values in parens when parsing an RE using the C++ wrapper."
1513    
1514    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1515        Characters greater than 255 were excluded from the class instead of being
1516        included.
1517    
1518    3.  The same bug as (2) above applied to negated POSIX classes such as
1519        [:^space:].
1520    
1521    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1522        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1523        I have changed it.
1524    
1525    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1526        first named subpattern) and a construct such as (?&a) would reference the
1527        first named subpattern whose name started with "a" (in other words, the
1528        length check was missing). Both these problems are fixed. "Subpattern name
1529        expected" is now given for (?&) (a zero-length name), and this patch also
1530        makes it give the same error for \k'' (previously it complained that that
1531        was a reference to a non-existent subpattern).
1532    
1533    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1534        this is right because (?- can be followed by option settings as well as by
1535        digits. I have, however, made the messages clearer.
1536    
1537    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1538        than the number used in the conditional) now cause a compile-time error.
1539        This is actually not compatible with Perl, which accepts such patterns, but
1540        treats the conditional as always being FALSE (as PCRE used to), but it
1541        seems to me that giving a diagnostic is better.
1542    
1543    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1544        and messages.
1545    
1546    9.  Fix two occurrences of "backslash" in comments that should have been
1547        "backspace".
1548    
1549    10. Remove two redundant lines of code that can never be obeyed (their function
1550        was moved elsewhere).
1551    
1552    11. The program that makes PCRE's Unicode character property table had a bug
1553        which caused it to generate incorrect table entries for sequences of
1554        characters that have the same character type, but are in different scripts.
1555        It amalgamated them into a single range, with the script of the first of
1556        them. In other words, some characters were in the wrong script. There were
1557        thirteen such cases, affecting characters in the following ranges:
1558    
1559          U+002b0 - U+002c1
1560          U+0060c - U+0060d
1561          U+0061e - U+00612
1562          U+0064b - U+0065e
1563          U+0074d - U+0076d
1564          U+01800 - U+01805
1565          U+01d00 - U+01d77
1566          U+01d9b - U+01dbf
1567          U+0200b - U+0200f
1568          U+030fc - U+030fe
1569          U+03260 - U+0327f
1570          U+0fb46 - U+0fbb1
1571          U+10450 - U+1049d
1572    
1573    12. The -o option (show only the matching part of a line) for pcregrep was not
1574        compatible with GNU grep in that, if there was more than one match in a
1575        line, it showed only the first of them. It now behaves in the same way as
1576        GNU grep.
1577    
1578    13. If the -o and -v options were combined for pcregrep, it printed a blank
1579        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1580        does the same. The return code can be used to tell if there were any
1581        non-matching lines.
1582    
1583    14. Added --file-offsets and --line-offsets to pcregrep.
1584    
1585    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1586        infinitely looping recursion. The bug was that positive lookaheads were not
1587        being skipped when checking for a possible empty match (negative lookaheads
1588        and both kinds of lookbehind were skipped).
1589    
1590    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1591        inclusion of <windows.h> to before rather than after the definition of
1592        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1593    
1594    17. Specifying a possessive quantifier with a specific limit for a Unicode
1595        character property caused pcre_compile() to compile bad code, which led at
1596        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1597        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1598        caused the error; without that there was no problem.
1599    
1600    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1601    
1602    19. Added --enable-pcretest-libreadline.
1603    
1604    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1605        RE::GlobalReplace(). As a result, the number of replacements returned was
1606        double what it should be. I removed one of the increments, but Craig sent a
1607        later patch that removed the other one (the right fix) and added unit tests
1608        that check the return values (which was not done before).
1609    
1610    21. Several CMake things:
1611    
1612        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1613            the names libpcre and libpcreposix, not just pcre and pcreposix.
1614    
1615        (2) The above change means that pcretest and pcregrep are now correctly
1616            linked with the newly-built libraries, not previously installed ones.
1617    
1618        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1619    
1620    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1621        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1622        UTF-8 newline character). The key issue is that the pattern starts .*;
1623        this means that the match must be either at the beginning, or after a
1624        newline. The bug was in the code for advancing after a failed match and
1625        checking that the new position followed a newline. It was not taking
1626        account of UTF-8 characters correctly.
1627    
1628    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1629        character classes. PCRE was not treating the sequence [:...:] as a
1630        character class unless the ... were all letters. Perl, however, seems to
1631        allow any characters between [: and :], though of course it rejects as
1632        unknown any "names" that contain non-letters, because all the known class
1633        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1634        for example, whereas PCRE did not - it did not recognize a POSIX character
1635        class. This seemed a bit dangerous, so the code has been changed to be
1636        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1637        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1638        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1639        Perl does, and where it didn't before.
1640    
1641    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1642        Windows environments %n is disabled by default.
1643    
1644    
1645    Version 7.4 21-Sep-07
1646    ---------------------
1647    
1648    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1649        means that a class such as [\s] counted as "explicit reference to CR or
1650        LF". That isn't really right - the whole point of the change was to try to
1651        help when there was an actual mention of one of the two characters. So now
1652        the change happens only if \r or \n (or a literal CR or LF) character is
1653        encountered.
1654    
1655    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1656        of both had grown to the point where there were only 3 bits left.
1657        Fortunately, there was spare space in the data structure, and so I have
1658        moved the internal flags into a new 16-bit field to free up more option
1659        bits.
1660    
1661    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1662        but did not set the internal JCHANGED flag - either of these is enough to
1663        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1664        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1665        start sets both bits.
1666    
1667    4.  Added options (at build time, compile time, exec time) to change \R from
1668        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1669    
1670    5.  doc/pcresyntax.html was missing from the distribution.
1671    
1672    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1673        compatibility, even though it is no longer used.
1674    
1675    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1676        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1677        windows.h file is present (where different names are used). [This was
1678        reversed later after testing - see 16 below.]
1679    
1680    8.  Changed all #include <config.h> to #include "config.h". There were also
1681        some further <pcre.h> cases that I changed to "pcre.h".
1682    
1683    9.  When pcregrep was used with the --colour option, it missed the line ending
1684        sequence off the lines that it output.
1685    
1686    10. It was pointed out to me that arrays of string pointers cause lots of
1687        relocations when a shared library is dynamically loaded. A technique of
1688        using a single long string with a table of offsets can drastically reduce
1689        these. I have refactored PCRE in four places to do this. The result is
1690        dramatic:
1691    
1692          Originally:                          290
1693          After changing UCP table:            187
1694          After changing error message table:   43
1695          After changing table of "verbs"       36
1696          After changing table of Posix names   22
1697    
1698        Thanks to the folks working on Gregex for glib for this insight.
1699    
1700    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1701        unicode-properties was also set.
1702    
1703    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1704    
1705    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1706        checked only for CRLF.
1707    
1708    14. Added casts to pcretest.c to avoid compiler warnings.
1709    
1710    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1711    
1712    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1713        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1714        entirely. This removes changes made in 7 above.
1715    
1716    17. The CMake files have been updated, and there is now more information about
1717        building with CMake in the NON-UNIX-USE document.
1718    
1719    
1720    Version 7.3 28-Aug-07
1721    ---------------------
1722    
1723     1. In the rejigging of the build system that eventually resulted in 7.1, the
1724        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1725        brackets there is not right, since it causes compilers to look for an
1726        installed pcre.h, not the version that is in the source that is being
1727        compiled (which of course may be different). I have changed it back to:
1728    
1729          #include "pcre.h"
1730    
1731        I have a vague recollection that the change was concerned with compiling in
1732        different directories, but in the new build system, that is taken care of
1733        by the VPATH setting the Makefile.
1734    
1735     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1736        when the subject happened to end in the byte 0x85 (e.g. if the last
1737        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1738        characters but of course it shouldn't be taken as a newline when it is part
1739        of another character. The bug was that, for an unlimited repeat of . in
1740        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1741        characters when looking for a newline.
1742    
1743     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1744    
1745     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1746        in debug output.
1747    
1748     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1749        long printing in the pcrecpp unittest when running under MinGW.
1750    
1751     6. ESC_K was left out of the EBCDIC table.
1752    
1753     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1754        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1755        limit also applies to "virtual nesting" when a pattern is recursive, and in
1756        this case 1000 isn't so big. I have been able to remove this limit at the
1757        expense of backing off one optimization in certain circumstances. Normally,
1758        when pcre_exec() would call its internal match() function recursively and
1759        immediately return the result unconditionally, it uses a "tail recursion"
1760        feature to save stack. However, when a subpattern that can match an empty
1761        string has an unlimited repetition quantifier, it no longer makes this
1762        optimization. That gives it a stack frame in which to save the data for
1763        checking that an empty string has been matched. Previously this was taken
1764        from the 1000-entry workspace that had been reserved. So now there is no
1765        explicit limit, but more stack is used.
1766    
1767     8. Applied Daniel's patches to solve problems with the import/export magic
1768        syntax that is required for Windows, and which was going wrong for the
1769        pcreposix and pcrecpp parts of the library. These were overlooked when this
1770        problem was solved for the main library.
1771    
1772     9. There were some crude static tests to avoid integer overflow when computing
1773        the size of patterns that contain repeated groups with explicit upper
1774        limits. As the maximum quantifier is 65535, the maximum group length was
1775        set at 30,000 so that the product of these two numbers did not overflow a
1776        32-bit integer. However, it turns out that people want to use groups that
1777        are longer than 30,000 bytes (though not repeat them that many times).
1778        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1779        made it possible to implement the integer overflow checks in a much more
1780        dynamic way, which I have now done. The artificial limitation on group
1781        length has been removed - we now have only the limit on the total length of
1782        the compiled pattern, which depends on the LINK_SIZE setting.
1783    
1784    10. Fixed a bug in the documentation for get/copy named substring when
1785        duplicate names are permitted. If none of the named substrings are set, the
1786        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1787        empty string.
1788    
1789    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1790        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1791        because the ] is interpreted as the first data character and the
1792        terminating ] is not found. PCRE has been made compatible with Perl in this
1793        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1794        cause memory overwriting.
1795    
1796    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1797        string has been matched (to stop an infinite loop). It was not recognizing
1798        a conditional subpattern that could match an empty string if that
1799        subpattern was within another subpattern. For example, it looped when
1800        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1801        condition was not nested. This bug has been fixed.
1802    
1803    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1804        past the start of the subject in the presence of bytes with the top bit
1805        set, for example "\x8aBCD".
1806    
1807    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1808        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1809    
1810    14. Optimized (?!) to (*FAIL).
1811    
1812    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1813        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1814        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1815        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1816        does: it's just the validity check that is more restrictive.
1817    
1818    16. Inserted checks for integer overflows during escape sequence (backslash)
1819        processing, and also fixed erroneous offset values for syntax errors during
1820        backslash processing.
1821    
1822    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1823        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1824    
1825    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1826        caused an overrun.
1827    
1828    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1829        something other than just ASCII characters) inside a group that had an
1830        unlimited repeat caused a loop at compile time (while checking to see
1831        whether the group could match an empty string).
1832    
1833    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1834        [\P{Any}] did so. (Error in the code for printing property names.)
1835    
1836    21. An orphan \E inside a character class could cause a crash.
1837    
1838    22. A repeated capturing bracket such as (A)? could cause a wild memory
1839        reference during compilation.
1840    
1841    23. There are several functions in pcre_compile() that scan along a compiled
1842        expression for various reasons (e.g. to see if it's fixed length for look
1843        behind). There were bugs in these functions when a repeated \p or \P was
1844        present in the pattern. These operators have additional parameters compared
1845        with \d, etc, and these were not being taken into account when moving along
1846        the compiled data. Specifically:
1847    
1848        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1849            length.
1850    
1851        (b) An item such as \pL+ within a repeated group could cause crashes or
1852            loops.
1853    
1854        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1855            "reference to non-existent subpattern" error.
1856    
1857        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1858    
1859    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1860        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1861    
1862    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1863    
1864    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1865        character were causing crashes (broken optimization).
1866    
1867    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1868        \p or \P) caused a compile-time loop.
1869    
1870    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1871        break. For example, the unstudied pattern [\r\n]A does not match the string
1872        "\r\nA" because change 7.0/46 below moves the current point on by two
1873        characters after failing to match at the start. However, the pattern \nA
1874        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1875        the same is true. There doesn't seem any very clean way out of this, but
1876        what I have chosen to do makes the common cases work: PCRE now takes note
1877        of whether there can be an explicit match for \r or \n anywhere in the
1878        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1879        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1880        pattern has explicit CR or LF references.
1881    
1882    29. Added (*CR) etc for changing newline setting at start of pattern.
1883    
1884    
1885    Version 7.2 19-Jun-07
1886    ---------------------
1887    
1888     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1889        which is apparently normally available under Windows.
1890    
1891     2. Re-jig the pcregrep tests with different newline settings in an attempt
1892        to make them independent of the local environment's newline setting.
1893    
1894     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1895    
1896     4. Some of the "internals" tests were previously cut out when the link size
1897        was not 2, because the output contained actual offsets. The recent new
1898        "Z" feature of pcretest means that these can be cut out, making the tests
1899        usable with all link sizes.
1900    
1901     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1902        stack recursion. This gives a massive performance boost under BSD, but just
1903        a small improvement under Linux. However, it saves one field in the frame
1904        in all cases.
1905    
1906     6. Added more features from the forthcoming Perl 5.10:
1907    
1908        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1909            recursion call. It refers to the nth most recently opened parentheses.
1910    
1911        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1912            to be opened parentheses.
1913    
1914        (c) Conditions that refer to capturing parentheses can be specified
1915            relatively, for example, (?(-2)... or (?(+3)...
1916    
1917        (d) \K resets the start of the current match so that everything before
1918            is not part of it.
1919    
1920        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1921    
1922        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1923            reference syntax.
1924    
1925        (g) (?| introduces a group in which the numbering of parentheses in each
1926            alternative starts with the same number.
1927    
1928        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1929    
1930     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1931        PCRE_INFO_JCHANGED.
1932    
1933     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1934        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1935        for detecting groups that can match an empty string.
1936    
1937     9. A pattern with a very large number of alternatives (more than several
1938        hundred) was running out of internal workspace during the pre-compile
1939        phase, where pcre_compile() figures out how much memory will be needed. A
1940        bit of new cunning has reduced the workspace needed for groups with
1941        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1942        workspace instead of running out of the 4096 that are available.
1943    
1944    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1945    
1946    11. Applied patch from Google to remove an optimization that didn't quite work.
1947        The report of the bug said:
1948    
1949          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1950          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1951          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1952    
1953    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1954        it matched the wrong number of bytes.
1955    
1956    
1957    Version 7.1 24-Apr-07
1958    ---------------------
1959    
1960     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1961        that is more "standard", making use of automake and other Autotools. There
1962        is some re-arrangement of the files and adjustment of comments consequent
1963        on this.
1964    
1965     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1966        for recursive directory scanning broke on some systems because the files
1967        are not scanned in any specific order and on different systems the order
1968        was different. A call to "sort" has been inserted into RunGrepTest for the
1969        approprate test as a short-term fix. In the longer term there may be an
1970        alternative.
1971    
1972     3. I had an email from Eric Raymond about problems translating some of PCRE's
1973        man pages to HTML (despite the fact that I distribute HTML pages, some
1974        people do their own conversions for various reasons). The problems
1975        concerned the use of low-level troff macros .br and .in. I have therefore
1976        removed all such uses from the man pages (some were redundant, some could
1977        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1978        HTML has been updated to handle .nf/.fi and to complain if it encounters
1979        .br or .in.
1980    
1981     4. Updated comments in configure.ac that get placed in config.h.in and also
1982        arranged for config.h to be included in the distribution, with the name
1983        config.h.generic, for the benefit of those who have to compile without
1984        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1985    
1986     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1987        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1988        makevp.bat for the current PCRE, using the additional files
1989        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1990    
1991     6. A Windows user reported a minor discrepancy with test 2, which turned out
1992        to be caused by a trailing space on an input line that had got lost in his
1993        copy. The trailing space was an accident, so I've just removed it.
1994    
1995     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1996        that is needed.
1997    
1998     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1999        as "const" (a) because they are and (b) because it helps the PHP
2000        maintainers who have recently made a script to detect big data structures
2001        in the php code that should be moved to the .rodata section. I remembered
2002        to update Builducptable as well, so it won't revert if ucptable.h is ever
2003        re-created.
2004    
2005     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2006        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2007        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2008        support is not required. This saves 1.5-2K of code, which is important in
2009        some applications.
2010    
2011        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2012        so as not to refer to the tables, even though these functions will never be
2013        called when UTF-8 support is disabled. Otherwise there are problems with a
2014        shared library.
2015    
2016    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2017    
2018        (a) It was defining its arguments as char * instead of void *.
2019    
2020        (b) It was assuming that all moves were upwards in memory; this was true
2021            a long time ago when I wrote it, but is no longer the case.
2022    
2023        The emulated memove() is provided for those environments that have neither
2024        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2025        is clearly not the case, as these two bugs were recently reported.
2026    
2027    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2028        and Detrail to create the HTML documentation, the .txt form of the man
2029        pages, and it removes trailing spaces from listed files. It also creates
2030        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2031        case, it wraps all the #defines with #ifndefs. This script should be run
2032        before "make dist".
2033    
2034    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2035        with Unicode property support.
2036    
2037        (a) For a maximizing quantifier, if the two different cases of the
2038            character were of different lengths in their UTF-8 codings (there are
2039            some cases like this - I found 11), and the matching function had to
2040            back up over a mixture of the two cases, it incorrectly assumed they
2041            were both the same length.
2042    
2043        (b) When PCRE was configured to use the heap rather than the stack for
2044            recursion during matching, it was not correctly preserving the data for
2045            the other case of a UTF-8 character when checking ahead for a match
2046            while processing a minimizing repeat. If the check also involved
2047            matching a wide character, but failed, corruption could cause an
2048            erroneous result when trying to check for a repeat of the original
2049            character.
2050    
2051    13. Some tidying changes to the testing mechanism:
2052    
2053        (a) The RunTest script now detects the internal link size and whether there
2054            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2055            values substituted by "configure". (The RunGrepTest script already did
2056            this for UTF-8.) The configure.ac script no longer substitutes the
2057            relevant variables.
2058    
2059        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2060            with length and offset values. This means that the output is different
2061            for different internal link sizes. Test 2 is skipped for link sizes
2062            other than 2 because of this, bypassing the problem. Unfortunately,
2063            there was also a test in test 3 (the locale tests) that used /B and
2064            failed for link sizes other than 2. Rather than cut the whole test out,
2065            I have added a new /Z option to pcretest that replaces the length and
2066            offset values with spaces. This is now used to make test 3 independent
2067            of link size. (Test 2 will be tidied up later.)
2068    
2069    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2070        segmentation fault instead of returning the appropriate error message.
2071    
2072    15. In multiline mode when the newline sequence was set to "any", the pattern
2073        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2074        This doesn't seem right; it now treats the CRLF combination as the line
2075        ending, and so does not match in that case. It's only a pattern such as ^$
2076        that would hit this one: something like ^ABC$ would have failed after \r
2077        and then tried again after \r\n.
2078    
2079    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2080        in an attempt to make files that differ only in their line terminators
2081        compare equal. This works on Linux.
2082    
2083    17. Under certain error circumstances pcregrep might try to free random memory
2084        as it exited. This is now fixed, thanks to valgrind.
2085    
2086    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2087        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2088        was because its rules for how to advance for /g after matching an empty
2089        string at the end of a line did not allow for this case. They now check for
2090        it specially.
2091    
2092    20. pcretest is supposed to handle patterns and data of any length, by
2093        extending its buffers when necessary. It was getting this wrong when the
2094        buffer for a data line had to be extended.
2095    
2096    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2097        CRLF as a newline sequence.
2098    
2099    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2100        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2101        I have nevertheless tidied it up.
2102    
2103    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2104    
2105    24. Added a man page for pcre-config.
2106    
2107    
2108    Version 7.0 19-Dec-06
2109    ---------------------
2110    
2111     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2112        moving to gcc 4.1.1.
2113    
2114     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2115        sys/time.h, which is documented as needed for this function. It doesn't
2116        seem to matter on Linux, but it showed up on some releases of OS X.
2117    
2118     3. It seems that there are systems where bytes whose values are greater than
2119        127 match isprint() in the "C" locale. The "C" locale should be the
2120        default when a C program starts up. In most systems, only ASCII printing
2121        characters match isprint(). This difference caused the output from pcretest
2122        to vary, making some of the tests fail. I have changed pcretest so that:
2123    
2124        (a) When it is outputting text in the compiled version of a pattern, bytes
2125            other than 32-126 are always shown as hex escapes.
2126    
2127        (b) When it is outputting text that is a matched part of a subject string,
2128            it does the same, unless a different locale has been set for the match
2129            (using the /L modifier). In this case, it uses isprint() to decide.
2130    
2131     4. Fixed a major bug that caused incorrect computation of the amount of memory
2132        required for a compiled pattern when options that changed within the
2133        pattern affected the logic of the preliminary scan that determines the
2134        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2135        that the computed length was too small. The symptoms of this bug were
2136        either the PCRE error "internal error: code overflow" from pcre_compile(),
2137        or a glibc crash with a message such as "pcretest: free(): invalid next
2138        size (fast)". Examples of patterns that provoked this bug (shown in
2139        pcretest format) are:
2140    
2141          /(?-x: )/x
2142          /(?x)(?-x: \s*#\s*)/
2143          /((?i)[\x{c0}])/8
2144          /(?i:[\x{c0}])/8
2145    
2146        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2147        is now done differently.
2148    
2149     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2150        wrapper classes; (b) implement a new function in the C++ scanner that is
2151        more efficient than the old way of doing things because it avoids levels of
2152        recursion in the regex matching; (c) add a paragraph to the documentation
2153        for the FullMatch() function.
2154    
2155     6. The escape sequence \n was being treated as whatever was defined as
2156        "newline". Not only was this contrary to the documentation, which states
2157        that \n is character 10 (hex 0A), but it also went horribly wrong when
2158        "newline" was defined as CRLF. This has been fixed.
2159    
2160     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2161        was being set to -1 for the "end of line" case (supposedly a value that no
2162        character can have). Though this value is never used (the check for end of
2163        line is "zero bytes in current character"), it caused compiler complaints.
2164        I've changed it to 0xffffffff.
2165    
2166     8. In pcre_version.c, the version string was being built by a sequence of
2167        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2168        string (as it is for production releases) called a macro with an empty
2169        argument. The C standard says the result of this is undefined. The gcc
2170        compiler treats it as an empty string (which was what was wanted) but it is
2171        reported that Visual C gives an error. The source has been hacked around to
2172        avoid this problem.
2173    
2174     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2175        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2176        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2177        of them did).
2178    
2179    10. Originally, pcretest opened its input and output without "b"; then I was
2180        told that "b" was needed in some environments, so it was added for release
2181        5.0 to both the input and output. (It makes no difference on Unix-like
2182        systems.) Later I was told that it is wrong for the input on Windows. I've
2183        now abstracted the modes into two macros, to make it easier to fiddle with
2184        them, and removed "b" from the input mode under Windows.
2185    
2186    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2187    
2188    12. Added -help and --help to pcretest as an official way of being reminded
2189        of the options.
2190    
2191    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2192        and pcrecpp.cc because they annoy compilers at high warning levels.
2193    
2194    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2195    
2196    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2197        scripts are not C programs :-) and which was not noticed because it works
2198        on Linux.
2199    
2200    16. pcretest is supposed to handle any length of pattern and data line (as one
2201        line or as a continued sequence of lines) by extending its input buffer if
2202        necessary. This feature was broken for very long pattern lines, leading to
2203        a string of junk being passed to pcre_compile() if the pattern was longer
2204        than about 50K.
2205    
2206    17. I have done a major re-factoring of the way pcre_compile() computes the
2207        amount of memory needed for a compiled pattern. Previously, there was code
2208        that made a preliminary scan of the pattern in order to do this. That was
2209        OK when PCRE was new, but as the facilities have expanded, it has become
2210        harder and harder to keep it in step with the real compile phase, and there
2211        have been a number of bugs (see for example, 4 above). I have now found a
2212        cunning way of running the real compile function in a "fake" mode that
2213        enables it to compute how much memory it would need, while actually only
2214        ever using a few hundred bytes of working memory and without too many
2215        tests of the mode. This should make future maintenance and development
2216        easier. A side effect of this work is that the limit of 200 on the nesting
2217        depth of parentheses has been removed (though this was never a serious
2218        limitation, I suspect). However, there is a downside: pcre_compile() now
2219        runs more slowly than before (30% or more, depending on the pattern). I
2220        hope this isn't a big issue. There is no effect on runtime performance.
2221    
2222    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2223        newline (only possible for the last line of a file) and it was a
2224        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2225    
2226    19. Added additional timing features to pcretest. (1) The -tm option now times
2227        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2228        separate command line item, by a number that specifies the number of
2229        repeats to use when timing. The default is 50000; this gives better
2230        precision, but takes uncomfortably long for very large patterns.
2231    
2232    20. Extended pcre_study() to be more clever in cases where a branch of a
2233        subpattern has no definite first character. For example, (a*|b*)[cd] would
2234        previously give no result from pcre_study(). Now it recognizes that the
2235        first character must be a, b, c, or d.
2236    
2237    21. There was an incorrect error "recursive call could loop indefinitely" if
2238        a subpattern (or the entire pattern) that was being tested for matching an
2239        empty string contained only one non-empty item after a nested subpattern.
2240        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2241        incorrectly, because the \d was being skipped in the check.
2242    
2243    22. The pcretest program now has a new pattern option /B and a command line
2244        option -b, which is equivalent to adding /B to every pattern. This causes
2245        it to show the compiled bytecode, without the additional information that
2246        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2247        is the same as /B/I).
2248    
2249    23. A new optimization is now able automatically to treat some sequences such
2250        as a*b as a*+b. More specifically, if something simple (such as a character
2251        or a simple class like \d) has an unlimited quantifier, and is followed by
2252        something that cannot possibly match the quantified thing, the quantifier
2253        is automatically "possessified".
2254    
2255    24. A recursive reference to a subpattern whose number was greater than 39
2256        went wrong under certain circumstances in UTF-8 mode. This bug could also
2257        have affected the operation of pcre_study().
2258    
2259    25. Realized that a little bit of performance could be had by replacing
2260        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2261    
2262    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2263    
2264    27. Possessive quantifiers such as a++ were previously implemented by turning
2265        them into atomic groups such as ($>a+). Now they have their own opcodes,
2266        which improves performance. This includes the automatically created ones
2267        from 23 above.
2268    
2269    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2270        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2271        the first matched character to be a colon. This applied both to named and
2272        numbered groups.
2273    
2274    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2275    
2276    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2277        building PCRE on Windows easier, so I have included it in the distribution.
2278    
2279    31. There is now a check in pcretest against a ridiculously large number being
2280        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2281        loop, the loop is abandoned.
2282    
2283    32. Forward references to subpatterns in conditions such as (?(2)...) where
2284        subpattern 2 is defined later cause pcre_compile() to search forwards in
2285        the pattern for the relevant set of parentheses. This search went wrong
2286        when there were unescaped parentheses in a character class, parentheses
2287        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2288    
2289    33. "Subroutine" calls and backreferences were previously restricted to
2290        referencing subpatterns earlier in the regex. This restriction has now
2291        been removed.
2292    
2293    34. Added a number of extra features that are going to be in Perl 5.10. On the
2294        whole, these are just syntactic alternatives for features that PCRE had
2295        previously implemented using the Python syntax or my own invention. The
2296        other formats are all retained for compatibility.
2297    
2298        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2299            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2300            also .NET compatible.
2301    
2302        (b) A recursion or subroutine call to a named group can now be defined as
2303            (?&name) as well as (?P>name).
2304    
2305        (c) A backreference to a named group can now be defined as \k<name> or
2306            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2307            5.10, are also .NET compatible.
2308    
2309        (d) A conditional reference to a named group can now use the syntax
2310            (?(<name>) or (?('name') as well as (?(name).
2311    
2312        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2313            groups (named and numbered) that are never evaluated inline, but can be
2314            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2315            is always false. There may be only one alternative in such a group.
2316    
2317        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2318            as the simple (?(R). The condition is true only if the most recent
2319            recursion is that of the given number or name. It does not search out
2320            through the entire recursion stack.
2321    
2322        (g) The escape \gN or \g{N} has been added, where N is a positive or
2323            negative number, specifying an absolute or relative reference.
2324    
2325    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2326        some "unreachable code" warnings.
2327    
2328    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2329        things, this adds five new scripts.
2330    
2331    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2332        There were also incompatibilities regarding the handling of \Q..\E inside
2333        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2334        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2335    
2336    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2337        matches an empty string, and forcibly breaks the loop. There were bugs in
2338        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2339        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2340        separate and independent bugs (that affected different cases) have been
2341        fixed.
2342    
2343    39. Refactored the code to abolish the use of different opcodes for small
2344        capturing bracket numbers. This is a tidy that I avoided doing when I
2345        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2346        The new approach is not only tidier, it makes it possible to reduce the
2347        memory needed to fix the previous bug (38).
2348    
2349    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2350        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2351        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2352        mode.
2353    
2354    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2355        report.
2356    
2357    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2358        copy construction and assignment in the C++ wrapper.
2359    
2360    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2361        couple of bugs that could have given wrong results in the "--newline=crlf"
2362        case.
2363    
2364    44. Added a number of casts and did some reorganization of signed/unsigned int
2365        variables following suggestions from Dair Grant. Also renamed the variable
2366        "this" as "item" because it is a C++ keyword.
2367    
2368    45. Arranged for dftables to add
2369    
2370          #include "pcre_internal.h"
2371    
2372        to pcre_chartables.c because without it, gcc 4.x may remove the array
2373        definition from the final binary if PCRE is built into a static library and
2374        dead code stripping is activated.
2375    
2376    46. For an unanchored pattern, if a match attempt fails at the start of a
2377        newline sequence, and the newline setting is CRLF or ANY, and the next two
2378        characters are CRLF, advance by two characters instead of one.
2379    
2380    
2381    Version 6.7 04-Jul-06
2382    ---------------------
2383    
2384     1. In order to handle tests when input lines are enormously long, pcretest has
2385        been re-factored so that it automatically extends its buffers when
2386        necessary. The code is crude, but this _is_ just a test program. The
2387        default size has been increased from 32K to 50K.
2388    
2389     2. The code in pcre_study() was using the value of the re argument before
2390        testing it for NULL. (Of course, in any sensible call of the function, it
2391        won't be NULL.)
2392    
2393     3. The memmove() emulation function in pcre_internal.h, which is used on
2394        systems that lack both memmove() and bcopy() - that is, hardly ever -
2395        was missing a "static" storage class specifier.
2396    
2397     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2398        containing an extended class (one that cannot be represented by a bitmap
2399        because it contains high-valued characters or Unicode property items, e.g.
2400        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2401        pattern, but PCRE should not loop if you do not (it no longer does).
2402        [Detail: two cases were found: (a) a repeated subpattern containing an
2403        extended class; (b) a recursive reference to a subpattern that followed a
2404        previous extended class. It wasn't skipping over the extended class
2405        correctly when UTF-8 mode was not set.]
2406    
2407     5. A negated single-character class was not being recognized as fixed-length
2408        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2409        compile error "lookbehind assertion is not fixed length".
2410    
2411     6. The RunPerlTest auxiliary script was showing an unexpected difference
2412        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2413        write a Perl script that can interpret lines of an input file either as
2414        byte characters or as UTF-8, which is what "perltest" was being required to
2415        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2416        can't do is switch easily at run time between having the "use utf8;" pragma
2417        or not. In the end, I fudged it by using the RunPerlTest script to insert
2418        "use utf8;" explicitly for the UTF-8 tests.
2419    
2420     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2421        the end of the subject string, contrary to the documentation and to what
2422        Perl does. This was true of both matching functions. Now it matches only at
2423        the start of the subject and immediately after *internal* newlines.
2424    
2425     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2426        a pointer to an int instead of a pointer to an unsigned long int. This
2427        caused problems on 64-bit systems.
2428    
2429     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2430        instance of the 'standard' template library not being so standard".
2431    
2432    10. There was no check on the number of named subpatterns nor the maximum
2433        length of a subpattern name. The product of these values is used to compute
2434        the size of the memory block for a compiled pattern. By supplying a very
2435        long subpattern name and a large number of named subpatterns, the size
2436        computation could be caused to overflow. This is now prevented by limiting
2437        the length of names to 32 characters, and the number of named subpatterns
2438        to 10,000.
2439    
2440    11. Subpatterns that are repeated with specific counts have to be replicated in
2441        the compiled pattern. The size of memory for this was computed from the
2442        length of the subpattern and the repeat count. The latter is limited to
2443        65535, but there was no limit on the former, meaning that integer overflow
2444        could in principle occur. The compiled length of a repeated subpattern is
2445        now limited to 30,000 bytes in order to prevent this.
2446    
2447    12. Added the optional facility to have named substrings with the same name.
2448    
2449    13. Added the ability to use a named substring as a condition, using the
2450        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2451        are numbers (not recommended). Forward references are permitted.
2452    
2453    14. Added forward references in named backreferences (if you see what I mean).
2454    
2455    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2456        pattern could run off the end of the subject. For example, the pattern
2457        "(?s)(.{1,5})"8 did this with the subject "ab".
2458    
2459    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2460        PCRE_CASELESS was set when matching characters that were quantified with ?
2461        or *.
2462    
2463    17. A character class other than a single negated character that had a minimum
2464        but no maximum quantifier - for example [ab]{6,} - was not handled
2465        correctly by pce_dfa_exec(). It would match only one character.
2466    
2467    18. A valid (though odd) pattern that looked like a POSIX character
2468        class but used an invalid character after [ (for example [[,abc,]]) caused
2469        pcre_compile() to give the error "Failed: internal error: code overflow" or
2470        in some cases to crash with a glibc free() error. This could even happen if
2471        the pattern terminated after [[ but there just happened to be a sequence of
2472        letters, a binary zero, and a closing ] in the memory that followed.
2473    
2474    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2475        over the years. Originally (before any Unicode support), just the bottom 8
2476        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2477        output from "man perlunicode" includes this:
2478    
2479          The regular expression compiler produces polymorphic opcodes.  That
2480          is, the pattern adapts to the data and automatically switches to
2481          the Unicode character scheme when presented with Unicode data--or
2482          instead uses a traditional byte scheme when presented with byte
2483          data.
2484    
2485        Sadly, a wide octal escape does not cause a switch, and in a string with
2486        no other multibyte characters, these octal escapes are treated as before.
2487        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2488        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2489        Unicode string.
2490    
2491        I have not perpetrated such confusion in PCRE. Up till now, it took just
2492        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2493        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2494        translate to the appropriate multibyte character.
2495    
2496    29. Applied some refactoring to reduce the number of warnings from Microsoft
2497        and Borland compilers. This has included removing the fudge introduced
2498        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2499        a warning about an unused variable.
2500    
2501    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2502        characters since release 4.0, because Perl (from release 5.004) does not.
2503        [Or at least, is documented not to: some releases seem to be in conflict
2504        with the documentation.] However, when a pattern was studied with
2505        pcre_study() and all its branches started with \s, PCRE still included VT
2506        as a possible starting character. Of course, this did no harm; it just
2507        caused an unnecessary match attempt.
2508    
2509    22. Removed a now-redundant internal flag bit that recorded the fact that case
2510        dependency changed within the pattern. This was once needed for "required
2511        byte" processing, but is no longer used. This recovers a now-scarce options
2512        bit. Also moved the least significant internal flag bit to the most-
2513        significant bit of the word, which was not previously used (hangover from
2514        the days when it was an int rather than a uint) to free up another bit for
2515        the future.
2516    
2517    23. Added support for CRLF line endings as well as CR and LF. As well as the
2518        default being selectable at build time, it can now be changed at runtime
2519        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2520        specify that it is scanning data with non-default line endings.
2521    
2522    24. Changed the definition of CXXLINK to make it agree with the definition of
2523        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2524    
2525    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2526        recursions. This makes a big different to stack usage for some patterns.
2527    
2528    26. If a subpattern containing a named recursion or subroutine reference such
2529        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2530        the space required for the compiled pattern went wrong and gave too small a
2531        value. Depending on the environment, this could lead to "Failed: internal
2532        error: code overflow at offset 49" or "glibc detected double free or
2533        corruption" errors.
2534    
2535    27. Applied patches from Google (a) to support the new newline modes and (b) to
2536        advance over multibyte UTF-8 characters in GlobalReplace.
2537    
2538    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2539        difference for some implementation of PCRE in some Windows version.
2540    
2541    29. Added some extra testing facilities to pcretest:
2542    
2543        \q<number>   in a data line sets the "match limit" value
2544        \Q<number>   in a data line sets the "match recursion limt" value
2545        -S <number>  sets the stack size, where <number> is in megabytes
2546    
2547        The -S option isn't available for Windows.
2548    
2549    
2550    Version 6.6 06-Feb-06
2551    ---------------------
2552    
2553     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2554        in pcreposix.h. I have copied the definition from pcre.h.
2555    
2556     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2557        because pcre.h is no longer a built file.
2558    
2559     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2560        not normally included in the compiled code.
2561    
2562    
2563    Version 6.5 01-Feb-06
2564    ---------------------
2565    
2566     1. When using the partial match feature with pcre_dfa_exec(), it was not
2567        anchoring the second and subsequent partial matches at the new starting
2568        point. This could lead to incorrect results. For example, with the pattern
2569        /1234/, partially matching against "123" and then "a4" gave a match.
2570    
2571     2. Changes to pcregrep:
2572    
2573        (a) All non-match returns from pcre_exec() were being treated as failures
2574            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2575            error message is output. Some extra information is given for the
2576            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2577            probably the only errors that are likely to be caused by users (by
2578            specifying a regex that has nested indefinite repeats, for instance).
2579            If there are more than 20 of these errors, pcregrep is abandoned.
2580    
2581        (b) A binary zero was treated as data while matching, but terminated the
2582            output line if it was written out. This has been fixed: binary zeroes
2583            are now no different to any other data bytes.
2584    
2585        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2586            used to set a locale for matching. The --locale=xxxx long option has
2587            been added (no short equivalent) to specify a locale explicitly on the
2588            pcregrep command, overriding the environment variables.
2589    
2590        (d) When -B was used with -n, some line numbers in the output were one less
2591            than they should have been.
2592    
2593        (e) Added the -o (--only-matching) option.
2594    
2595        (f) If -A or -C was used with -c (count only), some lines of context were
2596            accidentally printed for the final match.
2597    
2598        (g) Added the -H (--with-filename) option.
2599    
2600        (h) The combination of options -rh failed to suppress file names for files
2601            that were found from directory arguments.
2602    
2603        (i) Added the -D (--devices) and -d (--directories) options.
2604    
2605        (j) Added the -F (--fixed-strings) option.
2606    
2607        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2608    
2609        (l) Added the --colo(u)r option.
2610    
2611        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2612            is not present by default.
2613    
2614     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2615        items such as (?R) or (?1), when the recursion could match a number of
2616        alternatives. If it matched one of the alternatives, but subsequently,
2617        outside the recursion, there was a failure, the code tried to back up into
2618        the recursion. However, because of the way PCRE is implemented, this is not
2619        possible, and the result was an incorrect result from the match.
2620    
2621        In order to prevent this happening, the specification of recursion has
2622        been changed so that all such subpatterns are automatically treated as
2623        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2624    
2625     4. I had overlooked the fact that, in some locales, there are characters for
2626        which isalpha() is true but neither isupper() nor islower() are true. In
2627        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2628        and ordfeminine) are like this. This affected the treatment of \w and \W
2629        when they appeared in character classes, but not when they appeared outside
2630        a character class. The bit map for "word" characters is now created
2631        separately from the results of isalnum() instead of just taking it from the
2632        upper, lower, and digit maps. (Plus the underscore character, of course.)
2633    
2634     5. The above bug also affected the handling of POSIX character classes such as
2635        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2636        permanent tables. Instead, the bit maps for such a class were previously
2637        created as the appropriate unions of the upper, lower, and digit bitmaps.
2638        Now they are created by subtraction from the [[:word:]] class, which has
2639        its own bitmap.
2640    
2641     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2642        It is created by subtracting the vertical space characters (\x09, \x0a,
2643        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2644        subtraction was done in the overall bitmap for a character class, meaning
2645        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2646        be recognized. This bug has been fixed.
2647    
2648     7. Patches from the folks at Google:
2649    
2650          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2651          real life, but is still worth protecting against".
2652    
2653          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2654          regular expressions".
2655    
2656          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2657          have it.
2658    
2659          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2660          "configure" and the latter not, in order to fix a problem somebody had
2661          with compiling the Arg class on HP-UX.
2662    
2663          (e) Improve the error-handling of the C++ wrapper a little bit.
2664    
2665          (f) New tests for checking recursion limiting.
2666    
2667     8. The pcre_memmove() function, which is used only if the environment does not
2668        have a standard memmove() function (and is therefore rarely compiled),
2669        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2670        returning a result (though PCRE never actually uses the result).
2671    
2672     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2673        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2674        returned instead of calling malloc() with an overflowing number that would
2675        most likely cause subsequent chaos.
2676    
2677    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2678    
2679    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2680        with this option is matched, the nmatch and pmatch options of regexec() are
2681        ignored.
2682    
2683    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2684        provided in case anyone wants to the the POSIX interface with UTF-8
2685        strings.
2686    
2687    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2688        C++ linking (needed for some HP-UX environments).
2689    
2690    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2691        (unused parameter) and in the pcre_printint() function (omitted "default"
2692        switch label when the default is to do nothing).
2693    
2694    15. Added some code to make it possible, when PCRE is compiled as a C++
2695        library, to replace subject pointers for pcre_exec() with a smart pointer
2696        class, thus making it possible to process discontinuous strings.
2697    
2698    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2699        much the same function. They were added by different people who were trying
2700        to make PCRE easy to compile on non-Unix systems. It has been suggested
2701        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2702        for compiling on Windows systems. I have therefore replaced it with
2703        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2704        defaults to "extern" for C or "extern C" for C++, which works fine on
2705        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2706        SCOPE with something explicit in config.h. In addition:
2707    
2708        (a) pcreposix.h still had just "extern" instead of either of these macros;
2709            I have replaced it with PCRE_DATA_SCOPE.
2710    
2711        (b) Functions such as _pcre_xclass(), which are internal to the library,
2712            but external in the C sense, all had PCRE_EXPORT in their definitions.
2713            This is apparently wrong for the Windows case, so I have removed it.
2714            (It makes no difference on Unix-like systems.)
2715    
2716    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2717        of recursive calls to match(). This is different to MATCH_LIMIT because
2718        that limits the total number of calls to match(), not all of which increase
2719        the depth of recursion. Limiting the recursion depth limits the amount of
2720        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2721        when PCRE is compiled, and changed at run time. A patch from Google adds
2722        this functionality to the C++ interface.
2723    
2724    18. Changes to the handling of Unicode character properties:
2725    
2726        (a) Updated the table to Unicode 4.1.0.
2727    
2728        (b) Recognize characters that are not in the table as "Cn" (undefined).
2729    
2730        (c) I revised the way the table is implemented to a much improved format
2731            which includes recognition of ranges. It now supports the ranges that
2732            are defined in UnicodeData.txt, and it also amalgamates other
2733            characters into ranges. This has reduced the number of entries in the
2734            table from around 16,000 to around 3,000, thus reducing its size
2735            considerably. I realized I did not need to use a tree structure after
2736            all - a binary chop search is just as efficient. Having reduced the
2737            number of entries, I extended their size from 6 bytes to 8 bytes to
2738            allow for more data.
2739    
2740        (d) Added support for Unicode script names via properties such as \p{Han}.
2741    
2742    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2743        matching that character.
2744    
2745    20. When matching a repeated Unicode property with a minimum greater than zero,
2746        (for example \pL{2,}), PCRE could look past the end of the subject if it
2747        reached it while seeking the minimum number of characters. This could
2748        happen only if some of the characters were more than one byte long, because
2749        there is a check for at least the minimum number of bytes.
2750    
2751    21. Refactored the implementation of \p and \P so as to be more general, to
2752        allow for more different types of property in future. This has changed the
2753        compiled form incompatibly. Anybody with saved compiled patterns that use
2754        \p or \P will have to recompile them.
2755    
2756    22. Added "Any" and "L&" to the supported property types.
2757    
2758    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2759        but give a compile time error if the value is greater than 0xff.
2760    
2761    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2762        accidentally not being installed or uninstalled.
2763    
2764    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2765        made were to insert the current release number. This seemed silly, because
2766        it made things harder for people building PCRE on systems that don't run
2767        "configure". I have turned pcre.h into a distributed file, no longer built
2768        by "configure", with the version identification directly included. There is
2769        no longer a pcre.h.in file.
2770    
2771        However, this change necessitated a change to the pcre-config script as
2772        well. It is built from pcre-config.in, and one of the substitutions was the
2773        release number. I have updated configure.ac so that ./configure now finds
2774        the release number by grepping pcre.h.
2775    
2776    26. Added the ability to run the tests under valgrind.
2777    
2778    
2779    Version 6.4 05-Sep-05
2780    ---------------------
2781    
2782     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2783        "--" to be printed when multiple files were scanned, even when none of the
2784        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2785        consider it to be a bug, and have restored the previous behaviour.
2786    
2787     2. A couple of code tidies to get rid of compiler warnings.
2788    
2789     3. The pcretest program used to cheat by referring to symbols in the library
2790        whose names begin with _pcre_. These are internal symbols that are not
2791        really supposed to be visible externally, and in some environments it is
2792        possible to suppress them. The cheating is now confined to including
2793        certain files from the library's source, which is a bit cleaner.
2794    
2795     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2796        file's purpose clearer.
2797    
2798     5. Reorganized pcre_ucp_findchar().
2799    
2800    
2801    Version 6.3 15-Aug-05
2802    ---------------------
2803    
2804     1. The file libpcre.pc.in did not have general read permission in the tarball.
2805    
2806     2. There were some problems when building without C++ support:
2807    
2808        (a) If C++ support was not built, "make install" and "make test" still
2809            tried to test it.
2810    
2811        (b) There were problems when the value of CXX was explicitly set. Some
2812            changes have been made to try to fix these, and ...
2813    
2814        (c) --disable-cpp can now be used to explicitly disable C++ support.
2815    
2816        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2817            backslash in a target when C++ was disabled. This confuses some
2818            versions of "make", apparently. Using an intermediate variable solves
2819            this. (Same for CPP_LOBJ.)
2820    
2821     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2822        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2823        necessary on certain architectures.
2824    
2825     4. Added a setting of -export-symbols-regex to the link command to remove
2826        those symbols that are exported in the C sense, but actually are local
2827        within the library, and not documented. Their names all begin with
2828        "_pcre_". This is not a perfect job, because (a) we have to except some
2829        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2830        available (and never for static libraries). I have made a note to try to
2831        find a way round (a) in the future.
2832    
2833    
2834    Version 6.2 01-Aug-05
2835    ---------------------
2836    
2837     1. There was no test for integer overflow of quantifier values. A construction
2838        such as {1111111111111111} would give undefined results. What is worse, if
2839        a minimum quantifier for a parenthesized subpattern overflowed and became
2840        negative, the calculation of the memory size went wrong. This could have
2841        led to memory overwriting.
2842    
2843     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2844    
2845     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2846        operating environments where this matters.
2847    
2848     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2849        PCRE options from within the C++ wrapper.
2850    
2851     5. Named capturing subpatterns were not being correctly counted when a pattern
2852        was compiled. This caused two problems: (a) If there were more than 100
2853        such subpatterns, the calculation of the memory needed for the whole
2854        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2855        back references of the form \12, where the number was greater than 9, were
2856        not recognized as back references, even though there were sufficient
2857        previous subpatterns.
2858    
2859     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2860        versions of gcc, e.g. 2.95.4.
2861    
2862    
2863    Version 6.1 21-Jun-05
2864    ---------------------
2865    
2866     1. There was one reference to the variable "posix" in pcretest.c that was not
2867        surrounded by "#if !defined NOPOSIX".
2868    
2869     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2870        the cross-check on the old pcre_info() function, for the benefit of the
2871        cut-down version of PCRE that is currently imported into Exim.
2872    
2873     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2874        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2875        patterns that start (?i) but I don't think that matters. The use of (?i) is
2876        just an example; this all applies to the other options as well.
2877    
2878     4. Since libtool seems to echo the compile commands it is issuing, the output
2879        from "make" can be reduced a bit by putting "@" in front of each libtool
2880        compile command.
2881    
2882     5. Patch from the folks at Google for configure.in to be a bit more thorough
2883        in checking for a suitable C++ installation before trying to compile the
2884        C++ stuff. This should fix a reported problem when a compiler was present,
2885        but no suitable headers.
2886    
2887     6. The man pages all had just "PCRE" as their title. I have changed them to
2888        be the relevant file name. I have also arranged that these names are
2889        retained in the file doc/pcre.txt, which is a concatenation in text format
2890        of all the man pages except the little individual ones for each function.
2891    
2892     7. The NON-UNIX-USE file had not been updated for the different set of source
2893        files that come with release 6. I also added a few comments about the C++
2894        wrapper.
2895    
2896    
2897    Version 6.0 07-Jun-05
2898    ---------------------
2899    
2900     1. Some minor internal re-organization to help with my DFA experiments.
2901    
2902     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2903        didn't matter for the library itself when fully configured, but did matter
2904        when compiling without UCP support, or within Exim, where the ucp files are
2905        not imported.
2906    
2907     3. Refactoring of the library code to split up the various functions into
2908        different source modules. The addition of the new DFA matching code (see
2909        below) to a single monolithic source would have made it really too
2910        unwieldy, quite apart from causing all the code to be include in a
2911        statically linked application, when only some functions are used. This is
2912        relevant even without the DFA addition now that patterns can be compiled in
2913        one application and matched in another.
2914    
2915        The downside of splitting up is that there have to be some external
2916        functions and data tables that are used internally in different modules of
2917        the library but which are not part of the API. These have all had their
2918        names changed to start with "_pcre_" so that they are unlikely to clash
2919        with other external names.
2920    
2921     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2922        a different (DFA) algorithm. Although it is slower than the original
2923        function, it does have some advantages for certain types of matching
2924        problem.
2925    
2926     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2927        including restarting after a partial match.
2928    
2929     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2930        defined when compiling for Windows was sent to me. I have put it into the
2931        code, though I have no means of testing or verifying it.
2932    
2933     7. Added the pcre_refcount() auxiliary function.
2934    
2935     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2936        match before or at the first newline in the subject string. In pcretest,
2937        the /f option on a pattern can be used to set this.
2938    
2939     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2940        would behave wrongly. This has been present in PCRE since release 4.0.
2941    
2942    10. A number of changes to the pcregrep command:
2943    
2944        (a) Refactored how -x works; insert ^(...)$ instead of setting
2945            PCRE_ANCHORED and checking the length, in preparation for adding
2946            something similar for -w.
2947    
2948        (b) Added the -w (match as a word) option.
2949    
2950        (c) Refactored the way lines are read and buffered so as to have more
2951            than one at a time available.
2952    
2953        (d) Implemented a pcregrep test script.
2954    
2955        (e) Added the -M (multiline match) option. This allows patterns to match
2956            over several lines of the subject. The buffering ensures that at least
2957            8K, or the rest of the document (whichever is the shorter) is available
2958            for matching (and similarly the previous 8K for lookbehind assertions).
2959    
2960        (f) Changed the --help output so that it now says
2961    
2962              -w, --word-regex(p)
2963    
2964            instead of two lines, one with "regex" and the other with "regexp"
2965            because that confused at least one person since the short forms are the
2966            same. (This required a bit of code, as the output is generated
2967            automatically from a table. It wasn't just a text change.)
2968    
2969        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2970            option but starts with a hyphen. Could be a pattern or a path name
2971            starting with a hyphen, for instance.
2972    
2973        (h) "-" can be given as a file name to represent stdin.
2974    
2975        (i) When file names are being printed, "(standard input)" is used for
2976            the standard input, for compatibility with GNU grep. Previously
2977            "<stdin>" was used.
2978    
2979        (j) The option --label=xxx can be used to supply a name to be used for
2980            stdin when file names are being printed. There is no short form.
2981    
2982        (k) Re-factored the options decoding logic because we are going to add
2983            two more options that take data. Such options can now be given in four
2984            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2985    
2986        (l) Added the -A, -B, and -C options for requesting that lines of context
2987            around matches be printed.
2988    
2989        (m) Added the -L option to print the names of files that do not contain
2990            any matching lines, that is, the complement of -l.
2991    
2992        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2993            continue to scan other files.
2994    
2995        (o) The -s option was incorrectly implemented. For compatibility with other
2996            greps, it now suppresses the error message for a non-existent or non-
2997            accessible file (but not the return code). There is a new option called
2998            -q that suppresses the output of matching lines, which was what -s was
2999            previously doing.
3000    
3001        (p) Added --include and --exclude options to specify files for inclusion
3002            and exclusion when recursing.
3003    
3004    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3005        Hopefully, it now does.
3006    
3007    12. Missing cast in pcre_study().
3008    
3009    13. Added an "uninstall" target to the makefile.
3010    
3011    14. Replaced "extern" in the function prototypes in Makefile.in with
3012        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3013        world, but is set differently for Windows.
3014    
3015    15. Added a second compiling function called pcre_compile2(). The only
3016        difference is that it has an extra argument, which is a pointer to an
3017        integer error code. When there is a compile-time failure, this is set
3018        non-zero, in addition to the error test pointer being set to point to an
3019        error message. The new argument may be NULL if no error number is required
3020        (but then you may as well call pcre_compile(), which is now just a
3021        wrapper). This facility is provided because some applications need a
3022        numeric error indication, but it has also enabled me to tidy up the way
3023        compile-time errors are handled in the POSIX wrapper.
3024    
3025    16. Added VPATH=.libs to the makefile; this should help when building with one
3026        prefix path and installing with another. (Or so I'm told by someone who
3027        knows more about this stuff than I do.)
3028    
3029    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3030        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3031        match everything, including newlines. This is not POSIX-compatible, but
3032        somebody wanted the feature. From pcretest it can be activated by using
3033        both the P and the s flags.
3034    
3035    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3036    
3037    19. libpcre.pc was being incorrectly installed as executable.
3038    
3039    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3040        it now also looks for '\r' so that it will work unmodified on Windows.
3041    
3042    21. Added Google's contributed C++ wrapper to the distribution.
3043    
3044    22. Added some untidy missing memory free() calls in pcretest, to keep
3045        Electric Fence happy when testing.
3046    
3047    
3048    
3049    Version 5.0 13-Sep-04
3050    ---------------------
3051    
3052     1. Internal change: literal characters are no longer packed up into items
3053        containing multiple characters in a single byte-string. Each character
3054        is now matched using a separate opcode. However, there may be more than one
3055        byte in the character in UTF-8 mode.
3056    
3057     2. The pcre_callout_block structure has two new fields: pattern_position and
3058        next_item_length. These contain the offset in the pattern to the next match
3059        item, and its length, respectively.
3060    
3061     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3062        insertion of callouts before each pattern item. Added the /C option to
3063        pcretest to make use of this.
3064    
3065     4. On the advice of a Windows user, the lines
3066    
3067          #if defined(_WIN32) || defined(WIN32)
3068          _setmode( _fileno( stdout ), 0x8000 );
3069          #endif  /* defined(_WIN32) || defined(WIN32) */
3070    
3071        have been added to the source of pcretest. This apparently does useful
3072        magic in relation to line terminators.
3073    
3074     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3075        for the benefit of those environments where the "b" makes a difference.
3076    
3077     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3078        to know about it. I have put a hack into configure.in that adds in code
3079        to set GCC=yes if CC=icc. This seems to end up at a point in the
3080        generated configure script that is early enough to affect the setting of
3081        compiler options, which is what is needed, but I have no means of testing
3082        whether it really works. (The user who reported this had patched the
3083        generated configure script, which of course I cannot do.)
3084    
3085        LATER: After change 22 below (new libtool files), the configure script
3086        seems to know about icc (and also ecc). Therefore, I have commented out
3087        this hack in configure.in.
3088    
3089     7. Added support for pkg-config (2 patches were sent in).
3090    
3091     8. Negated POSIX character classes that used a combination of internal tables
3092        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3093        [[:^ascii]]. Typically, they would match almost any characters. The other
3094        POSIX classes were not broken in this way.
3095    
3096     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3097        to find the match, as PCRE was deluded into thinking that the match had to
3098        start at the start point or following a newline. The same bug applied to
3099        patterns with negative forward assertions or any backward assertions
3100        preceding ".*" at the start, unless the pattern required a fixed first
3101        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3102    
3103    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3104        starting at the last subject character, bytes beyond the end of the subject
3105        string were read.
3106    
3107    11. Renamed the variable "class" as "classbits" to make life easier for C++
3108        users. (Previously there was a macro definition, but it apparently wasn't
3109        enough.)
3110    
3111    12. Added the new field "tables" to the extra data so that tables can be passed
3112        in at exec time, or the internal tables can be re-selected. This allows
3113        a compiled regex to be saved and re-used at a later time by a different
3114        program that might have everything at different addresses.
3115    
3116    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3117        -R library as well as a -L library.
3118    
3119    14. The debugging options of pcretest (-d on the command line or D on a
3120        pattern) showed incorrect output for anything following an extended class
3121        that contained multibyte characters and which was followed by a quantifier.
3122    
3123    15. Added optional support for general category Unicode character properties
3124        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3125        support. It adds about 90K to the size of the library. The meanings of the
3126        inbuilt class escapes such as \d and \s have NOT been changed.
3127    
3128    16. Updated pcredemo.c to include calls to free() to release the memory for the
3129        compiled pattern.
3130    
3131    17. The generated file chartables.c was being created in the source directory
3132        instead of in the building directory. This caused the build to fail if the
3133        source directory was different from the building directory, and was
3134        read-only.
3135    
3136    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3137        file. No doubt somebody will tell me if they don't make sense... Also added
3138        Dan Mooney's comments about building on OpenVMS.
3139    
3140    19. Added support for partial matching via the PCRE_PARTIAL option for
3141        pcre_exec() and the \P data escape in pcretest.
3142    
3143    20. Extended pcretest with 3 new pattern features:
3144    
3145        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3146              write the compiled pattern to the file whose name is "rest-of-line".
3147              This is a straight binary dump of the data, with the saved pointer to
3148              the character tables forced to be NULL. The study data, if any, is
3149              written too. After writing, pcretest reads a new pattern.
3150    
3151        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3152              compiled pattern from the given file. There must not be any
3153              occurrences of "<" in the file name (pretty unlikely); if there are,
3154              pcretest will instead treat the initial "<" as a pattern delimiter.
3155              After reading in the pattern, pcretest goes on to read data lines as
3156              usual.
3157    
3158        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3159              and 16-bit fields in a compiled pattern, to simulate a pattern that
3160              was compiled on a host of opposite endianness.
3161    
3162    21. The pcre-exec() function can now cope with patterns that were compiled on
3163        hosts of opposite endianness, with this restriction:
3164    
3165          As for any compiled expression that is saved and used later, the tables
3166          pointer field cannot be preserved; the extra_data field in the arguments
3167          to pcre_exec() should be used to pass in a tables address if a value
3168          other than the default internal tables were used at compile time.
3169    
3170    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3171        now diagnosed as an error. Previously, most of the time, a negative number
3172        would have been treated as zero, but if in addition "ovector" was passed as
3173        NULL, a crash could occur.
3174    
3175    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3176        new versions from the libtool 1.5 distribution (the last one is a copy of
3177        a file called libtool.m4). This seems to have fixed the need to patch
3178        "configure" to support Darwin 1.3 (which I used to do). However, I still
3179        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3180        workstation).
3181    
3182    24. Changed the PCRE licence to be the more standard "BSD" licence.
3183    
3184    
3185    Version 4.5 01-Dec-03
3186    ---------------------
3187    
3188     1. There has been some re-arrangement of the code for the match() function so
3189        that it can be compiled in a version that does not call itself recursively.
3190        Instead, it keeps those local variables that need separate instances for
3191        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3192        needs to "recurse". Keeping track of where control must go is done by means
3193        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3194        hide most of the details from the main code, and operates only if
3195        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3196        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3197        operating.
3198    
3199        To make it easier for callers to provide specially tailored get/free
3200        functions for this usage, two new functions, pcre_stack_malloc, and
3201        pcre_stack_free, are used. They are always called in strict stacking order,
3202        and the size of block requested is always the same.
3203    
3204        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3205        PCRE has been compiled to use the stack or the heap for recursion. The
3206        -C option of pcretest uses this to show which version is compiled.
3207    
3208        A new data escape \S, is added to pcretest; it causes the amounts of store
3209        obtained and freed by both kinds of malloc/free at match time to be added
3210        to the output.
3211    
3212     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3213        what's available on my current Linux desktop machine.
3214    
3215     3. When matching a UTF-8 string, the test for a valid string at the start has
3216        been extended. If start_offset is not zero, PCRE now checks that it points
3217        to a byte that is the start of a UTF-8 character. If not, it returns
3218        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3219        this is necessary because there may be backward assertions in the pattern.
3220        When matching the same subject several times, it may save resources to use
3221        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3222    
3223     4. The code for checking the validity of UTF-8 strings has been tightened so
3224        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3225        containing "overlong sequences".
3226    
3227     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3228        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3229        should have been "&", but it just so happened that all the cases this let
3230        through by mistake were picked up later in the function.
3231    
3232     6. I had used a variable called "isblank" - this is a C99 function, causing
3233        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3234    
3235     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3236        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3237        so the version is shown; (c) stop "make test" echoing "./RunTest".
3238    
3239     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3240    
3241     9. The prototype for memmove() for systems that don't have it was using
3242        size_t, but the inclusion of the header that defines size_t was later. I've
3243        moved the #includes for the C headers earlier to avoid this.
3244    
3245    10. Added some adjustments to the code to make it easier to compiler on certain
3246        special systems:
3247    
3248          (a) Some "const" qualifiers were missing.
3249          (b) Added the macro EXPORT before all exported functions; by default this
3250              is defined to be empty.
3251          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3252              that it reads its output file name as an argument instead of writing
3253              to the standard output and assuming this can be redirected.
3254    
3255    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3256        class containing characters with values greater than 255, PCRE compilation
3257        went into a loop.
3258    
3259    12. A recursive reference to a subpattern that was within another subpattern
3260        that had a minimum quantifier of zero caused PCRE to crash. For example,
3261        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3262        recursion. If the recursively-called subpattern itself had a zero repeat,
3263        that was OK.
3264    
3265    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3266        buffer into which it was copied (for escape processing) was still set at
3267        1024, so long lines caused crashes.
3268    
3269    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3270        "internal error: code overflow...". This applied to any character class
3271        that was followed by a possessive quantifier.
3272    
3273    15. Modified the Makefile to add libpcre.la as a prerequisite for
3274        libpcreposix.la because I was told this is needed for a parallel build to
3275        work.
3276    
3277    16. If a pattern that contained .* following optional items at the start was
3278        studied, the wrong optimizing data was generated, leading to matching
3279        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3280        matching string must start with a or b or c. The correct conclusion for
3281        this pattern is that a match can start with any character.
3282    
3283    
3284    Version 4.4 13-Aug-03
3285    ---------------------
3286    
3287     1. In UTF-8 mode, a character class containing characters with values between
3288        127 and 255 was not handled correctly if the compiled pattern was studied.
3289        In fixing this, I have also improved the studying algorithm for such
3290        classes (slightly).
3291    
3292     2. Three internal functions had redundant arguments passed to them. Removal
3293        might give a very teeny performance improvement.
3294    
3295     3. Documentation bug: the value of the capture_top field in a callout is *one
3296        more than* the number of the hightest numbered captured substring.
3297    
3298     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3299        in incorrectly linking with a previously installed version. They now link
3300        explicitly with libpcre.la.
3301    
3302     5. configure.in no longer needs to recognize Cygwin specially.
3303    
3304     6. A problem in pcre.in for Windows platforms is fixed.
3305    
3306     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3307        pcretest, it used to include the size of the study block as part of its
3308        output. Unfortunately, the structure contains a field that has a different
3309        size on different hardware architectures. This meant that the tests that
3310        showed this size failed. As the block is currently always of a fixed size,
3311        this information isn't actually particularly useful in pcretest output, so
3312        I have just removed it.
3313    
3314     8. Three pre-processor statements accidentally did not start in column 1.
3315        Sadly, there are *still* compilers around that complain, even though
3316        standard C has not required this for well over a decade. Sigh.
3317    
3318     9. In pcretest, the code for checking callouts passed small integers in the
3319        callout_data field, which is a void * field. However, some picky compilers
3320        complained about the casts involved for this on 64-bit systems. Now
3321        pcretest passes the address of the small integer instead, which should get
3322        rid of the warnings.
3323    
3324    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3325        both compile and run time, and gives an error if an invalid UTF-8 sequence
3326        is found. There is a option for disabling this check in cases where the
3327        string is known to be correct and/or the maximum performance is wanted.
3328    
3329    11. In response to a bug report, I changed one line in Makefile.in from
3330    
3331            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3332        to
3333            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3334    
3335        to look similar to other lines, but I have no way of telling whether this
3336        is the right thing to do, as I do not use Windows. No doubt I'll get told
3337        if it's wrong...
3338    
3339    
3340    Version 4.3 21-May-03
3341    ---------------------
3342    
3343    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3344       Makefile.
3345    
3346    2. Some refactoring to improve the quality of the code:
3347    
3348       (i)   The utf8_table... variables are now declared "const".
3349    
3350       (ii)  The code for \cx, which used the "case flipping" table to upper case
3351             lower case letters, now just substracts 32. This is ASCII-specific,
3352             but the whole concept of \cx is ASCII-specific, so it seems
3353             reasonable.
3354    
3355       (iii) PCRE was using its character types table to recognize decimal and
3356             hexadecimal digits in the pattern. This is silly, because it handles
3357             only 0-9, a-f, and A-F, but the character types table is locale-
3358             specific, which means strange things might happen. A private
3359             table is now used for this - though it costs 256 bytes, a table is
3360             much faster than multiple explicit tests. Of course, the standard
3361             character types table is still used for matching digits in subject
3362             strings against \d.
3363    
3364       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3365             ending in _t are). So I've renamed it as ESC_tee.
3366    
3367    3. The first argument for regexec() in the POSIX wrapper should have been
3368       defined as "const".
3369    
3370    4. Changed pcretest to use malloc() for its buffers so that they can be
3371       Electric Fenced for debugging.
3372    
3373    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3374       to read one or more bytes before the start of the subject string. Often this
3375       had no effect on PCRE's behaviour, but in some circumstances it could
3376       provoke a segmentation fault.
3377    
3378    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3379       to try to read one or more bytes before the start of the subject string.
3380    
3381    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3382       UTF-8 support could misbehave in various ways if the subject string
3383       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3384       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3385       back over UTF-8 characters.)
3386    
3387    
3388    Version 4.2 14-Apr-03
3389    ---------------------
3390    
3391    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3392    
3393    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3394         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3395         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3396         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3397         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3398           and BUILD_EXEEXT
3399         Note: automatic setting of the BUILD variables is not yet working
3400         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3401           compile-time but not at link-time
3402         [LINK]: use for linking executables only
3403         make different versions for Windows and non-Windows
3404         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3405           libraries
3406         [LINK_FOR_BUILD]: new variable
3407         [OBJEXT]: use throughout
3408         [EXEEXT]: use throughout
3409         <winshared>: new target
3410         <wininstall>: new target
3411         <dftables.o>: use native compiler
3412         <dftables>: use native linker
3413         <install>: handle Windows platform correctly
3414         <clean>: ditto
3415         <check>: ditto
3416         copy DLL to top builddir before testing
3417    
3418       As part of these changes, -no-undefined was removed again. This was reported
3419       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3420       in any case.
3421    
3422    3. Some tidies to get rid of compiler warnings:
3423    
3424       . In the match_data structure, match_limit was an unsigned long int, whereas
3425         match_call_count was an int. I've made them both unsigned long ints.
3426    
3427       . In pcretest the fact that a const uschar * doesn't automatically cast to
3428         a void * provoked a warning.
3429    
3430       . Turning on some more compiler warnings threw up some "shadow" variables
3431         and a few more missing casts.
3432    
3433    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3434       option, a class that contained a single character with a value between 128
3435       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3436    
3437    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3438       option, a class that contained several characters, but with at least one
3439       whose value was between 128 and 255 caused PCRE to crash.
3440    
3441    
3442    Version 4.1 12-Mar-03
3443    ---------------------
3444    
3445    1. Compiling with gcc -pedantic found a couple of places where casts were
3446    needed, and a string in dftables.c that was longer than standard compilers are
3447    required to support.
3448    
3449    2. Compiling with Sun's compiler found a few more places where the code could
3450    be tidied up in order to avoid warnings.
3451    
3452    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3453    first of these names is deprecated in the latest Autoconf in favour of the name
3454    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3455    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3456    analogy I have changed it to CFLAGS_FOR_BUILD.
3457    
3458    4. Added -no-undefined to the linking command in the Makefile, because this is
3459    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3460    linking step for the pcreposix library.
3461    
3462    5. PCRE was failing to diagnose the case of two named groups with the same
3463    name.
3464    
3465    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3466    literal character that is needed in the subject for a match, and scans along to
3467    ensure that it is present before embarking on the full matching process. This
3468    saves time in cases of nested unlimited repeats that are never going to match.
3469    Problem: the scan can take a lot of time if the subject is very long (e.g.
3470    megabytes), thus penalizing straightforward matches. It is now done only if the
3471    amount of subject to be scanned is less than 1000 bytes.
3472    
3473    7. A lesser problem with the same optimization is that it was recording the
3474    first character of an anchored pattern as "needed", thus provoking a search
3475    right along the subject, even when the first match of the pattern was going to
3476    fail. The "needed" character is now not set for anchored patterns, unless it
3477    follows something in the pattern that is of non-fixed length. Thus, it still
3478    fulfils its original purpose of finding quick non-matches in cases of nested
3479    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3480    
3481    
3482    Version 4.0 17-Feb-03
3483    ---------------------
3484    
3485    1. If a comment in an extended regex that started immediately after a meta-item
3486    extended to the end of string, PCRE compiled incorrect data. This could lead to
3487    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3488    
3489    2. Moved to autoconf 2.53 and libtool 1.4.2.
3490    
3491    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3492    the special perltest8 script is no longer needed - all the tests can be run
3493    from a single perltest script.
3494    
3495    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3496    by \s. It has now been removed in PCRE. This means it isn't recognized as
3497    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3498    class [:space:] *does* include VT, thereby creating a mess.
3499    
3500    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3501    space and tab.
3502    
3503    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3504    its new features into the main test script, reducing the number of scripts.
3505    
3506    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3507    were backward compatible, and made the (?i) apply to the whole pattern, as if
3508    /i were given. Now it behaves more logically, and applies the option setting
3509    only to what follows. PCRE has been changed to follow suit. However, if it
3510    finds options settings right at the start of the pattern, it extracts them into
3511    the global options, as before. Thus, they show up in the info data.
3512    
3513    8. Added support for the \Q...\E escape sequence. Characters in between are
3514    treated as literals. This is slightly different from Perl in that $ and @ are
3515    also handled as literals inside the quotes. In Perl, they will cause variable
3516    interpolation. Note the following examples:
3517    
3518        Pattern            PCRE matches      Perl matches
3519    
3520        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3521        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3522        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3523    
3524    For compatibility with Perl, \Q...\E sequences are recognized inside character
3525    classes as well as outside them.
3526    
3527    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3528    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3529    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3530    signed/unsigned warnings.
3531    
3532    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3533    option for pcretest, so I've replaced it by a simple function that does just
3534    that job.
3535    
3536    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3537    "pcregrep -".
3538    
3539    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3540    Java package. This provides some syntactic sugar for simple cases of what my
3541    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3542    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3543    item, you can use this simplified notation. Note that only makes sense with
3544    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3545    greediness, whatever the setting of the PCRE_UNGREEDY option.
3546    
3547    13. A change of greediness default within a pattern was not taking effect at
3548    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3549    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3550    was abstracted outside.
3551    
3552    14. PCRE now supports the \G assertion. It is true when the current matching
3553    position is at the start point of the match. This differs from \A when the
3554    starting offset is non-zero. Used with the /g option of pcretest (or similar
3555    code), it works in the same way as it does for Perl's /g option. If all
3556    alternatives of a regex begin with \G, the expression is anchored to the start
3557    match position, and the "anchored" flag is set in the compiled expression.
3558    
3559    15. Some bugs concerning the handling of certain option changes within patterns
3560    have been fixed. These applied to options other than (?ims). For example,
3561    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3562    been the other way round. Some of this was related to change 7 above.
3563    
3564    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3565    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3566    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3567    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3568    
3569    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3570    mode. Unlike ".", it always matches newline, whatever the setting of
3571    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3572    assertions. Perl allows it, but it doesn't (in general) work because it can't
3573    calculate the length of the lookbehind. At least, that's the case for Perl
3574    5.8.0 - I've been told they are going to document that it doesn't work in
3575    future.
3576    
3577    18. Added an error diagnosis for escapes that PCRE does not support: these are
3578    \L, \l, \N, \P, \p, \U, \u, and \X.
3579    
3580    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3581    reading past the end of the pattern in cases such as /[abcd/.
3582    
3583    20. PCRE was getting more memory than necessary for patterns with classes that
3584    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3585    
3586    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3587    compiling PCRE for use with Virtual Pascal.
3588    
3589    22. Small fix to the Makefile to make it work properly if the build is done
3590    outside the source tree.
3591    
3592    23. Added a new extension: a condition to go with recursion. If a conditional
3593    subpattern starts with (?(R) the "true" branch is used if recursion has
3594    happened, whereas the "false" branch is used only at the top level.
3595    
3596    24. When there was a very long string of literal characters (over 255 bytes
3597    without UTF support, over 250 bytes with UTF support), the computation of how
3598    much memory was required could be incorrect, leading to segfaults or other
3599    strange effects.
3600    
3601    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3602    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3603    there was a subsequent back reference to those brackets. This meant that, for
3604    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3605    possible to check for precisely this case. All we can do is abandon the
3606    optimization if .* occurs inside capturing brackets when there are any back
3607    references whatsoever. (See below for a better fix that came later.)
3608    
3609    26. The handling of the optimization for finding the first character of a
3610    non-anchored pattern, and for finding a character that is required later in the
3611