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

Diff of /code/trunk/ChangeLog

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

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