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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 777 by zherczeg, Thu Dec 1 15:15:31 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  Version 8.21
5    ------------
6    
7    1.  Updating the JIT compiler.
8    
9    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10        are added as well.
11    
12    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14        calling _pcre_jit_exec. Some extra comments are added.
15    
16    4.  Mark settings inside atomic groups that do not contain any capturing
17        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18        was introduced by change 18 for 8.20.
19    
20    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21        ECMA-262 standard.
22    
23    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25        This bug was probably introduced by change 9 of 8.13.
26    
27    7.  While fixing 6 above, I noticed that a number of other items were being
28        incorrectly rejected as "not fixed length". This arose partly because newer
29        opcodes had not been added to the fixed-length checking code. I have (a)
30        corrected the bug and added tests for these items, and (b) arranged for an
31        error to occur if an unknown opcode is encountered while checking for fixed
32        length instead of just assuming "not fixed length". The items that were
33        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36    
37    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
38        being incorrectly compiled and would have given unpredicatble results.
39    
40    9.  A possessively repeated subpattern with minimum repeat count greater than
41        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
42        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
43        the first (A) could occur when it should not.
44    
45    10. Add a cast and remove a redundant test from the code.
46    
47    11. JIT should use pcre_malloc/pcre_free for allocation.
48    
49    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
50        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
51        is anything other than /usr, -L is still shown).
52    
53    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
54    
55    14. Perl does not support \N without a following name in a [] class; PCRE now
56        also gives an error.
57    
58    15. If a forward reference was repeated with an upper limit of around 2000,
59        it caused the error "internal error: overran compiling workspace". The
60        maximum number of forward references (including repeats) was limited by the
61        internal workspace, and dependent on the LINK_SIZE. The code has been
62        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
63        the default depends on LINK_SIZE. There is a new upper limit (for safety)
64        of around 200,000 forward references. While doing this, I also speeded up
65        the filling in of repeated forward references.
66    
67    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
68        incorrectly expecting the subject to contain another "a" after the start.
69    
70    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
71        in the match, the SKIP should be ignored. This was not happening; instead
72        the SKIP was being treated as NOMATCH. For patterns such as
73        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
74        tested.
75    
76    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
77        now much more compatible with Perl, in particular in cases where the result
78        is a non-match for a non-anchored pattern. For example, if
79        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
80        "m", where previously it did not return a name. A side effect of this
81        change is that for partial matches, the last encountered mark name is
82        returned, as for non matches. A number of tests that were previously not
83        Perl-compatible have been moved into the Perl-compatible test files. The
84        refactoring has had the pleasing side effect of removing one argument from
85        the match() function, thus reducing its stack requirements.
86    
87    19. If the /S+ option was used in pcretest to study a pattern using JIT,
88        subsequent uses of /S (without +) incorrectly behaved like /S+.
89    
90    21. Retrieve executable code size support for the JIT compiler and fixing
91        some warnings.
92    
93    
94    Version 8.20 21-Oct-2011
95    ------------------------
96    
97    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
98        a POSIX class. After further experiments with Perl, which convinced me that
99        Perl has bugs and confusions, a closing square bracket is no longer allowed
100        in a POSIX name. This bug also affected patterns with classes that started
101        with full stops.
102    
103    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
104        captured substring, but while checking the failing first alternative,
105        substring 1 is temporarily captured. If the output vector supplied to
106        pcre_exec() was not big enough for this capture, the yield of the function
107        was still zero ("insufficient space for captured substrings"). This cannot
108        be totally fixed without adding another stack variable, which seems a lot
109        of expense for a edge case. However, I have improved the situation in cases
110        such as /(a)(b)x|abc/ matched against "abc", where the return code
111        indicates that fewer than the maximum number of slots in the ovector have
112        been set.
113    
114    3.  Related to (2) above: when there are more back references in a pattern than
115        slots in the output vector, pcre_exec() uses temporary memory during
116        matching, and copies in the captures as far as possible afterwards. It was
117        using the entire output vector, but this conflicts with the specification
118        that only 2/3 is used for passing back captured substrings. Now it uses
119        only the first 2/3, for compatibility. This is, of course, another edge
120        case.
121    
122    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
123        main code base, and can be used by building with --enable-jit. When this is
124        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
125        runtime --no-jit option is given.
126    
127    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
128        ovector, the return from the function was zero, implying that there were
129        other matches that did not fit. The correct "exactly full" value is now
130        returned.
131    
132    6.  If a subpattern that was called recursively or as a subroutine contained
133        (*PRUNE) or any other control that caused it to give a non-standard return,
134        invalid errors such as "Error -26 (nested recursion at the same subject
135        position)" or even infinite loops could occur.
136    
137    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
138        computing the minimum length on reaching *ACCEPT, and so ended up with the
139        wrong value of 1 rather than 0. Further investigation indicates that
140        computing a minimum subject length in the presence of *ACCEPT is difficult
141        (think back references, subroutine calls), and so I have changed the code
142        so that no minimum is registered for a pattern that contains *ACCEPT.
143    
144    8.  If (*THEN) was present in the first (true) branch of a conditional group,
145        it was not handled as intended. [But see 16 below.]
146    
147    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
148        Sheri Pierce.
149    
150    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
151        the first byte in a match must be "a".
152    
153    11. Change 17 for 8.13 increased the recursion depth for patterns like
154        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
155        pattern contains any instances of (*THEN). If it does not, the old
156        optimizations are restored. It would be nice to do this on a per-group
157        basis, but at the moment that is not feasible.
158    
159    12. In some environments, the output of pcretest -C is CRLF terminated. This
160        broke RunTest's code that checks for the link size. A single white space
161        character after the value is now allowed for.
162    
163    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
164        For "fr", it uses the Windows-specific input and output files.
165    
166    14. If (*THEN) appeared in a group that was called recursively or as a
167        subroutine, it did not work as intended. [But see next item.]
168    
169    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
170        pattern fragments (but not containing any | characters). If A and B are
171        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
172        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
173        D. In other words, Perl considers parentheses that do not contain any |
174        characters to be part of a surrounding alternative, whereas PCRE was
175        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
176        differently. PCRE now behaves in the same way as Perl, except in the case
177        of subroutine/recursion calls such as (?1) which have in any case always
178        been different (but PCRE had them first :-).
179    
180    16. Related to 15 above: Perl does not treat the | in a conditional group as
181        creating alternatives. Such a group is treated in the same way as an
182        ordinary group without any | characters when processing (*THEN). PCRE has
183        been changed to match Perl's behaviour.
184    
185    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
186        RunGrepTest script failed.
187    
188    18. Change 22 for version 13 caused atomic groups to use more stack. This is
189        inevitable for groups that contain captures, but it can lead to a lot of
190        stack use in large patterns. The old behaviour has been restored for atomic
191        groups that do not contain any capturing parentheses.
192    
193    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
194        suppress the check for a minimum subject length at run time. (If it was
195        given to pcre_exec() or pcre_dfa_exec() it did work.)
196    
197    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
198        fail to work when decoding hex characters in data strings in EBCDIC
199        environments.
200    
201    21. It appears that in at least one Mac OS environment, the isxdigit() function
202        is implemented as a macro that evaluates to its argument more than once,
203        contravening the C 90 Standard (I haven't checked a later standard). There
204        was an instance in pcretest which caused it to go wrong when processing
205        \x{...} escapes in subject strings. The has been rewritten to avoid using
206        things like p++ in the argument of isxdigit().
207    
208    
209    Version 8.13 16-Aug-2011
210    ------------------------
211    
212    1.  The Unicode data tables have been updated to Unicode 6.0.0.
213    
214    2.  Two minor typos in pcre_internal.h have been fixed.
215    
216    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
217        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
218        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
219    
220    4.  There were a number of related bugs in the code for matching backrefences
221        caselessly in UTF-8 mode when codes for the characters concerned were
222        different numbers of bytes. For example, U+023A and U+2C65 are an upper
223        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
224        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
225        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
226        2-byte code at the end of the subject (it thought there wasn't enough data
227        left).
228    
229    5.  Comprehensive information about what went wrong is now returned by
230        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
231        as the output vector has at least 2 elements. The offset of the start of
232        the failing character and a reason code are placed in the vector.
233    
234    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
235        now returned is for the first byte of the failing character, instead of the
236        last byte inspected. This is an incompatible change, but I hope it is small
237        enough not to be a problem. It makes the returned offset consistent with
238        pcre_exec() and pcre_dfa_exec().
239    
240    7.  pcretest now gives a text phrase as well as the error number when
241        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
242        failure, the offset and reason code are output.
243    
244    8.  When \R was used with a maximizing quantifier it failed to skip backwards
245        over a \r\n pair if the subsequent match failed. Instead, it just skipped
246        back over a single character (\n). This seems wrong (because it treated the
247        two characters as a single entity when going forwards), conflicts with the
248        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
249        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
250        has been changed.
251    
252    9.  Some internal refactoring has changed the processing so that the handling
253        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
254        time (the PCRE_DOTALL option was changed this way some time ago: version
255        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
256        which was always a bit of a fudge. It also means that there is one less
257        argument for the match() function, which reduces its stack requirements
258        slightly. This change also fixes an incompatibility with Perl: the pattern
259        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
260    
261    10. More internal refactoring has drastically reduced the number of recursive
262        calls to match() for possessively repeated groups such as (abc)++ when
263        using pcre_exec().
264    
265    11. While implementing 10, a number of bugs in the handling of groups were
266        discovered and fixed:
267    
268        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
269        (a|)*(?1) gave a compile-time internal error.
270        ((a|)+)+  did not notice that the outer group could match an empty string.
271        (^a|^)+   was not marked as anchored.
272        (.*a|.*)+ was not marked as matching at start or after a newline.
273    
274    12. Yet more internal refactoring has removed another argument from the match()
275        function. Special calls to this function are now indicated by setting a
276        value in a variable in the "match data" data block.
277    
278    13. Be more explicit in pcre_study() instead of relying on "default" for
279        opcodes that mean there is no starting character; this means that when new
280        ones are added and accidentally left out of pcre_study(), testing should
281        pick them up.
282    
283    14. The -s option of pcretest has been documented for ages as being an old
284        synonym of -m (show memory usage). I have changed it to mean "force study
285        for every regex", that is, assume /S for every regex. This is similar to -i
286        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
287        using it. It makes it easier to run collections of tests with and without
288        study enabled, and thereby test pcre_study() more easily. All the standard
289        tests are now run with and without -s (but some patterns can be marked as
290        "never study" - see 20 below).
291    
292    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
293        restoration of the capturing data to the outer values was not happening
294        correctly.
295    
296    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
297        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
298        pattern had matched an empty string, and so incorrectly returned a no
299        match.
300    
301    17. There was optimizing code for the last branch of non-capturing parentheses,
302        and also for the obeyed branch of a conditional subexpression, which used
303        tail recursion to cut down on stack usage. Unfortunately, now that there is
304        the possibility of (*THEN) occurring in these branches, tail recursion is
305        no longer possible because the return has to be checked for (*THEN). These
306        two optimizations have therefore been removed. [But see 8.20/11 above.]
307    
308    18. If a pattern containing \R was studied, it was assumed that \R always
309        matched two bytes, thus causing the minimum subject length to be
310        incorrectly computed because \R can also match just one byte.
311    
312    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
313        was incorrectly computed.
314    
315    20. If /S is present twice on a test pattern in pcretest input, it now
316        *disables* studying, thereby overriding the use of -s on the command line
317        (see 14 above). This is necessary for one or two tests to keep the output
318        identical in both cases.
319    
320    21. When (*ACCEPT) was used in an assertion that matched an empty string and
321        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
322    
323    22. When an atomic group that contained a capturing parenthesis was
324        successfully matched, but the branch in which it appeared failed, the
325        capturing was not being forgotten if a higher numbered group was later
326        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
327        group 1 to "a", when in fact it should be unset. This applied to multi-
328        branched capturing and non-capturing groups, repeated or not, and also to
329        positive assertions (capturing in negative assertions does not happen
330        in PCRE) and also to nested atomic groups.
331    
332    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
333        subject after a captured substring, to make it easier to tell which of a
334        number of identical substrings has been captured.
335    
336    24. The way atomic groups are processed by pcre_exec() has been changed so that
337        if they are repeated, backtracking one repetition now resets captured
338        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
339        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
340        "aaa". Previously, it would have been "a". As part of this code
341        refactoring, the way recursive calls are handled has also been changed.
342    
343    25. If an assertion condition captured any substrings, they were not passed
344        back unless some other capturing happened later. For example, if
345        (?(?=(a))a) was matched against "a", no capturing was returned.
346    
347    26. When studying a pattern that contained subroutine calls or assertions,
348        the code for finding the minimum length of a possible match was handling
349        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
350        group 1 called group 2 while simultaneously a separate group 2 called group
351        1). A stack overflow occurred in this case. I have fixed this by limiting
352        the recursion depth to 10.
353    
354    27. Updated RunTest.bat in the distribution to the version supplied by Tom
355        Fortmann. This supports explicit test numbers on the command line, and has
356        argument validation and error reporting.
357    
358    28. An instance of \X with an unlimited repeat could fail if at any point the
359        first character it looked at was a mark character.
360    
361    29. Some minor code refactoring concerning Unicode properties and scripts
362        should reduce the stack requirement of match() slightly.
363    
364    30. Added the '=' option to pcretest to check the setting of unused capturing
365        slots at the end of the pattern, which are documented as being -1, but are
366        not included in the return count.
367    
368    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
369        compiled something random. Now it gives a compile-time error (as does
370        Perl).
371    
372    32. A *MARK encountered during the processing of a positive assertion is now
373        recorded and passed back (compatible with Perl).
374    
375    33. If --only-matching or --colour was set on a pcregrep call whose pattern
376        had alternative anchored branches, the search for a second match in a line
377        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
378        matched the line "0102" twice. The same bug affected patterns that started
379        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
380        twice.
381    
382    34. Previously, PCRE did not allow quantification of assertions. However, Perl
383        does, and because of capturing effects, quantifying parenthesized
384        assertions may at times be useful. Quantifiers are now allowed for
385        parenthesized assertions.
386    
387    35. A minor code tidy in pcre_compile() when checking options for \R usage.
388    
389    36. \g was being checked for fancy things in a character class, when it should
390        just be a literal "g".
391    
392    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
393        appearance of a nested POSIX class supersedes an apparent external class.
394        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
395        unescaped square brackets may also appear as part of class names. For
396        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
397        more like Perl. (But see 8.20/1 above.)
398    
399    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
400        was because it thought it was \N{name}, which is not supported).
401    
402    39. Add minix to OS list not supporting the -S option in pcretest.
403    
404    40. PCRE tries to detect cases of infinite recursion at compile time, but it
405        cannot analyze patterns in sufficient detail to catch mutual recursions
406        such as ((?1))((?2)). There is now a runtime test that gives an error if a
407        subgroup is called recursively as a subpattern for a second time at the
408        same position in the subject string. In previous releases this might have
409        been caught by the recursion limit, or it might have run out of stack.
410    
411    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
412        happen only once. PCRE was, however incorrectly giving a compile time error
413        "recursive call could loop indefinitely" because it cannot analyze the
414        pattern in sufficient detail. The compile time test no longer happens when
415        PCRE is compiling a conditional subpattern, but actual runaway loops are
416        now caught at runtime (see 40 above).
417    
418    42. It seems that Perl allows any characters other than a closing parenthesis
419        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
420        has been changed to be the same.
421    
422    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
423        as not to get warnings when autogen.sh is called. Also changed
424        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
425    
426    44. To help people who use pcregrep to scan files containing exceedingly long
427        lines, the following changes have been made:
428    
429        (a) The default value of the buffer size parameter has been increased from
430            8K to 20K. (The actual buffer used is three times this size.)
431    
432        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
433            PCRE is built.
434    
435        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
436            to be set at run time.
437    
438        (d) Numerical values in pcregrep options can be followed by K or M, for
439            example --buffer-size=50K.
440    
441        (e) If a line being scanned overflows pcregrep's buffer, an error is now
442            given and the return code is set to 2.
443    
444    45. Add a pointer to the latest mark to the callout data block.
445    
446    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
447        partial match of an empty string instead of no match. This was specific to
448        the use of ".".
449    
450    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
451        complete match instead of a partial match. This bug was dependent on both
452        the PCRE_UTF8 and PCRE_DOTALL options being set.
453    
454    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
455        starting byte set, because \b was not being ignored.
456    
457    
458    Version 8.12 15-Jan-2011
459    ------------------------
460    
461    1.  Fixed some typos in the markup of the man pages, and wrote a script that
462        checks for such things as part of the documentation building process.
463    
464    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
465        --match-limit and --recursion-limit options (added for 8.11). In
466        particular, this made one of the standard tests fail. (The integer value
467        went into the wrong half of a long int.)
468    
469    3.  If the --colour option was given to pcregrep with -v (invert match), it
470        did strange things, either producing crazy output, or crashing. It should,
471        of course, ignore a request for colour when reporting lines that do not
472        match.
473    
474    4.  Another pcregrep bug caused similar problems if --colour was specified with
475        -M (multiline) and the pattern match finished with a line ending.
476    
477    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
478        matched in multiline mode, the following line was shown as part of the
479        match. This seems wrong, so I have changed it.
480    
481    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
482        the check for further matches in the same line (so they could be coloured)
483        to overrun the end of the current line. If another match was found, it was
484        incorrectly shown (and then shown again when found in the next line).
485    
486    7.  If pcregrep was compiled under Windows, there was a reference to the
487        function pcregrep_exit() before it was defined. I am assuming this was
488        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
489        reported by a user. I've moved the definition above the reference.
490    
491    
492    Version 8.11 10-Dec-2010
493    ------------------------
494    
495    1.  (*THEN) was not working properly if there were untried alternatives prior
496        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
497        backtracked to try for "b" instead of moving to the next alternative branch
498        at the same level (in this case, to look for "c"). The Perl documentation
499        is clear that when (*THEN) is backtracked onto, it goes to the "next
500        alternative in the innermost enclosing group".
501    
502    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
503        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
504        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
505        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
506        (*THEN).
507    
508    3.  If \s appeared in a character class, it removed the VT character from
509        the class, even if it had been included by some previous item, for example
510        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
511        of \s, but is part of the POSIX "space" class.)
512    
513    4.  A partial match never returns an empty string (because you can always
514        match an empty string at the end of the subject); however the checking for
515        an empty string was starting at the "start of match" point. This has been
516        changed to the "earliest inspected character" point, because the returned
517        data for a partial match starts at this character. This means that, for
518        example, /(?<=abc)def/ gives a partial match for the subject "abc"
519        (previously it gave "no match").
520    
521    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
522        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
523        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
524        has an implication that the given string is incomplete (because a partial
525        match is preferred over a full match). For this reason, these items now
526        give a partial match in this situation. [Aside: previously, the one case
527        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
528        match rather than a full match, which was wrong by the old rules, but is
529        now correct.]
530    
531    6.  There was a bug in the handling of #-introduced comments, recognized when
532        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
533        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
534        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
535        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
536        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
537        places in pcre_compile().
538    
539    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
540        comments when looking ahead for named forward references to subpatterns,
541        the only newline sequence it recognized was NL. It now handles newlines
542        according to the set newline convention.
543    
544    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
545        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
546        cater for a lack of strerror(). These oversights have been fixed.
547    
548    9.  Added --match-limit and --recursion-limit to pcregrep.
549    
550    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
551    
552    11. When the -o option was used, pcregrep was setting a return code of 1, even
553        when matches were found, and --line-buffered was not being honoured.
554    
555    12. Added an optional parentheses number to the -o and --only-matching options
556        of pcregrep.
557    
558    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
559        can match an empty string. The code to do it in pcretest and pcredemo
560        needed fixing:
561    
562        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
563            only one byte after an empty string match just before CRLF (this case
564            just got forgotten; "any" and "anycrlf" were OK).
565    
566        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
567            mode when an empty string match preceded an ASCII character followed by
568            a non-ASCII character. (The code for advancing by one character rather
569            than one byte was nonsense.)
570    
571        (c) The pcredemo.c sample program did not have any code at all to handle
572            the cases when CRLF is a valid newline sequence.
573    
574    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
575        as a starting offset was within the subject string. There is now a new
576        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
577        negative or greater than the length of the string. In order to test this,
578        pcretest is extended to allow the setting of negative starting offsets.
579    
580    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
581        starting offset points to the beginning of a UTF-8 character was
582        unnecessarily clumsy. I tidied it up.
583    
584    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
585        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
586    
587    17. Nobody had reported that the --include_dir option, which was added in
588        release 7.7 should have been called --include-dir (hyphen, not underscore)
589        for compatibility with GNU grep. I have changed it to --include-dir, but
590        left --include_dir as an undocumented synonym, and the same for
591        --exclude-dir, though that is not available in GNU grep, at least as of
592        release 2.5.4.
593    
594    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
595        characters from a string of bytes) have been redefined so as not to use
596        loops, in order to improve performance in some environments. At the same
597        time, I abstracted some of the common code into auxiliary macros to save
598        repetition (this should not affect the compiled code).
599    
600    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
601        compile-time error is now given if \c is not followed by an ASCII
602        character, that is, a byte less than 128. (In EBCDIC mode, the code is
603        different, and any byte value is allowed.)
604    
605    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
606        START_OPTIMIZE option, which is now allowed at compile time - but just
607        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
608        to pcregrep and other applications that have no direct access to PCRE
609        options. The new /Y option in pcretest sets this option when calling
610        pcre_compile().
611    
612    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
613        back references. Groups containing recursive back references were forced to
614        be atomic by that change, but in the case of named groups, the amount of
615        memory required was incorrectly computed, leading to "Failed: internal
616        error: code overflow". This has been fixed.
617    
618    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
619        pcretest.c, to avoid build problems in some Borland environments.
620    
621    
622    Version 8.10 25-Jun-2010
623    ------------------------
624    
625    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
626        THEN.
627    
628    2.  (*ACCEPT) was not working when inside an atomic group.
629    
630    3.  Inside a character class, \B is treated as a literal by default, but
631        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
632        causes the error). The code is unchanged, but I tidied the documentation.
633    
634    4.  Inside a character class, PCRE always treated \R and \X as literals,
635        whereas Perl faults them if its -w option is set. I have changed PCRE so
636        that it faults them when PCRE_EXTRA is set.
637    
638    5.  Added support for \N, which always matches any character other than
639        newline. (It is the same as "." when PCRE_DOTALL is not set.)
640    
641    6.  When compiling pcregrep with newer versions of gcc which may have
642        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
643        declared with attribute warn_unused_result" were given. Just casting the
644        result to (void) does not stop the warnings; a more elaborate fudge is
645        needed. I've used a macro to implement this.
646    
647    7.  Minor change to pcretest.c to avoid a compiler warning.
648    
649    8.  Added four artifical Unicode properties to help with an option to make
650        \s etc use properties (see next item). The new properties are: Xan
651        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
652    
653    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
654        use Unicode properties. (*UCP) at the start of a pattern can be used to set
655        this option. Modified pcretest to add /W to test this facility. Added
656        REG_UCP to make it available via the POSIX interface.
657    
658    10. Added --line-buffered to pcregrep.
659    
660    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
661        studied, and the match started with a letter with a code point greater than
662        127 whose first byte was different to the first byte of the other case of
663        the letter, the other case of this starting letter was not recognized
664        (#976).
665    
666    12. If a pattern that was studied started with a repeated Unicode property
667        test, for example, \p{Nd}+, there was the theoretical possibility of
668        setting up an incorrect bitmap of starting bytes, but fortunately it could
669        not have actually happened in practice until change 8 above was made (it
670        added property types that matched character-matching opcodes).
671    
672    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
673        possible starting bytes for non-anchored patterns.
674    
675    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
676        \R, and also a number of cases that involve Unicode properties, both
677        explicit and implicit when PCRE_UCP is set.
678    
679    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
680        input, it could crash or give wrong results if characters with values
681        greater than 0xc0 were present in the subject string. (Detail: it assumed
682        UTF-8 input when processing these items.)
683    
684    16. Added a lot of (int) casts to avoid compiler warnings in systems where
685        size_t is 64-bit (#991).
686    
687    17. Added a check for running out of memory when PCRE is compiled with
688        --disable-stack-for-recursion (#990).
689    
690    18. If the last data line in a file for pcretest does not have a newline on
691        the end, a newline was missing in the output.
692    
693    19. The default pcre_chartables.c file recognizes only ASCII characters (values
694        less than 128) in its various bitmaps. However, there is a facility for
695        generating tables according to the current locale when PCRE is compiled. It
696        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
697        characters, are recognized by isspace() and therefore were getting set in
698        these tables, and indeed these tables seem to approximate to ISO 8859. This
699        caused a problem in UTF-8 mode when pcre_study() was used to create a list
700        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
701        which of course cannot start UTF-8 characters. I have changed the code so
702        that only real ASCII characters (less than 128) and the correct starting
703        bytes for UTF-8 encodings are set for characters greater than 127 when in
704        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
705        altogether.)
706    
707    20. Added the /T option to pcretest so as to be able to run tests with non-
708        standard character tables, thus making it possible to include the tests
709        used for 19 above in the standard set of tests.
710    
711    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
712        reference to a subpattern the other side of a comment that contains an
713        opening parenthesis caused either an internal compiling error, or a
714        reference to the wrong subpattern.
715    
716    
717    Version 8.02 19-Mar-2010
718    ------------------------
719    
720    1.  The Unicode data tables have been updated to Unicode 5.2.0.
721    
722    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
723        configured.
724    
725    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
726        original author of that file, following a query about its status.
727    
728    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
729        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
730    
731    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
732        quantifier applied to a forward-referencing subroutine call, could compile
733        incorrect code or give the error "internal error: previously-checked
734        referenced subpattern not found".
735    
736    6.  Both MS Visual Studio and Symbian OS have problems with initializing
737        variables to point to external functions. For these systems, therefore,
738        pcre_malloc etc. are now initialized to local functions that call the
739        relevant global functions.
740    
741    7.  There were two entries missing in the vectors called coptable and poptable
742        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
743        I've fixed the data, and added a kludgy way of testing at compile time that
744        the lengths are correct (equal to the number of opcodes).
745    
746    8.  Following on from 7, I added a similar kludge to check the length of the
747        eint vector in pcreposix.c.
748    
749    9.  Error texts for pcre_compile() are held as one long string to avoid too
750        much relocation at load time. To find a text, the string is searched,
751        counting zeros. There was no check for running off the end of the string,
752        which could happen if a new error number was added without updating the
753        string.
754    
755    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
756    
757    11. \K was not working if it appeared in an atomic group or in a group that
758        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
759        \K is "not well defined" if used in an assertion. PCRE now accepts it if
760        the assertion is positive, but not if it is negative.
761    
762    12. Change 11 fortuitously reduced the size of the stack frame used in the
763        "match()" function of pcre_exec.c by one pointer. Forthcoming
764        implementation of support for (*MARK) will need an extra pointer on the
765        stack; I have reserved it now, so that the stack frame size does not
766        decrease.
767    
768    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
769        item in branch that calls a recursion is a subroutine call - as in the
770        second branch in the above example - was incorrectly given the compile-
771        time error "recursive call could loop indefinitely" because pcre_compile()
772        was not correctly checking the subroutine for matching a non-empty string.
773    
774    14. The checks for overrunning compiling workspace could trigger after an
775        overrun had occurred. This is a "should never occur" error, but it can be
776        triggered by pathological patterns such as hundreds of nested parentheses.
777        The checks now trigger 100 bytes before the end of the workspace.
778    
779    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
780    
781    
782    Version 8.01 19-Jan-2010
783    ------------------------
784    
785    1.  If a pattern contained a conditional subpattern with only one branch (in
786        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
787        computed the wrong minimum data length (which is of course zero for such
788        subpatterns). This could cause incorrect "no match" results.
789    
790    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
791        the pattern is reset in the first branch, pcre_compile() failed with
792        "internal error: code overflow at offset...". This happened only when
793        the reset was to the original external option setting. (An optimization
794        abstracts leading options settings into an external setting, which was the
795        cause of this.)
796    
797    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
798        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
799        assertion pattern did not match (meaning that the assertion was true), it
800        was incorrectly treated as false if the SKIP had been reached during the
801        matching. This also applied to assertions used as conditions.
802    
803    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
804        assertion subpattern, including such a pattern used as a condition,
805        unpredictable results occurred, instead of the error return
806        PCRE_ERROR_DFA_UITEM.
807    
808    5.  The C++ GlobalReplace function was not working like Perl for the special
809        situation when an empty string is matched. It now does the fancy magic
810        stuff that is necessary.
811    
812    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
813        removed. (These were left over from very, very early versions of PCRE.)
814    
815    7.  Some cosmetic changes to the code to make life easier when compiling it
816        as part of something else:
817    
818        (a) Change DEBUG to PCRE_DEBUG.
819    
820        (b) In pcre_compile(), rename the member of the "branch_chain" structure
821            called "current" as "current_branch", to prevent a collision with the
822            Linux macro when compiled as a kernel module.
823    
824        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
825            prevent a collision with the Linux macro when compiled as a kernel
826            module.
827    
828    8.  In pcre_compile() there are some checks for integer overflows that used to
829        cast potentially large values to (double). This has been changed to that
830        when building, a check for int64_t is made, and if it is found, it is used
831        instead, thus avoiding the use of floating point arithmetic. (There is no
832        other use of FP in PCRE.) If int64_t is not found, the fallback is to
833        double.
834    
835    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
836        2005 (difference between two addresses compared to an unsigned value).
837    
838    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
839        custom one, because of the following reported problem in Windows:
840    
841          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
842              under Win32.
843          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
844              therefore missing the function definition.
845          - The compiler thus generates a "C" signature for the test function.
846          - The linker fails to find the "C" function.
847          - PCRE fails to configure if asked to do so against libbz2.
848    
849    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
850        messages were output:
851    
852          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
853          rerunning libtoolize, to keep the correct libtool macros in-tree.
854          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
855    
856        I have done both of these things.
857    
858    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
859        most of the time, it *can* run out if it is given a pattern that contains a
860        runaway infinite recursion. I updated the discussion in the pcrestack man
861        page.
862    
863    13. Now that we have gone to the x.xx style of version numbers, the minor
864        version may start with zero. Using 08 or 09 is a bad idea because users
865        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
866        interpreted as invalid octal numbers. I've updated the previous comment in
867        configure.ac, and also added a check that gives an error if 08 or 09 are
868        used.
869    
870    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
871        causing partial matching to fail when the end of the subject matched \W
872        in a UTF-8 pattern where \W was quantified with a minimum of 3.
873    
874    15. There were some discrepancies between the declarations in pcre_internal.h
875        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
876        their definitions. The declarations used "const uschar *" and the
877        definitions used USPTR. Even though USPTR is normally defined as "const
878        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
879        reported that: "This difference in casting confuses some C++ compilers, for
880        example, SunCC recognizes above declarations as different functions and
881        generates broken code for hbpcre." I have changed the declarations to use
882        USPTR.
883    
884    16. GNU libtool is named differently on some systems. The autogen.sh script now
885        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
886        (FreeBSD).
887    
888    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
889        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
890        comment: "Figure out how to create a longlong from a string: strtoll and
891        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
892        instance, but it only takes 2 args instead of 3!"
893    
894    18. A subtle bug concerned with back references has been fixed by a change of
895        specification, with a corresponding code fix. A pattern such as
896        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
897        refers, was giving matches when it shouldn't. For example, xa=xaaa would
898        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
899        same bug. Such groups have to be quantified to be useful, or contained
900        inside another quantified group. (If there's no repetition, the reference
901        can never match.) The problem arises because, having left the group and
902        moved on to the rest of the pattern, a later failure that backtracks into
903        the group uses the captured value from the final iteration of the group
904        rather than the correct earlier one. I have fixed this in PCRE by forcing
905        any group that contains a reference to itself to be an atomic group; that
906        is, there cannot be any backtracking into it once it has completed. This is
907        similar to recursive and subroutine calls.
908    
909    
910    Version 8.00 19-Oct-09
911    ----------------------
912    
913    1.  The table for translating pcre_compile() error codes into POSIX error codes
914        was out-of-date, and there was no check on the pcre_compile() error code
915        being within the table. This could lead to an OK return being given in
916        error.
917    
918    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
919        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
920        in a Windows environment.
921    
922    3.  The pcregrep --count option prints the count for each file even when it is
923        zero, as does GNU grep. However, pcregrep was also printing all files when
924        --files-with-matches was added. Now, when both options are given, it prints
925        counts only for those files that have at least one match. (GNU grep just
926        prints the file name in this circumstance, but including the count seems
927        more useful - otherwise, why use --count?) Also ensured that the
928        combination -clh just lists non-zero counts, with no names.
929    
930    4.  The long form of the pcregrep -F option was incorrectly implemented as
931        --fixed_strings instead of --fixed-strings. This is an incompatible change,
932        but it seems right to fix it, and I didn't think it was worth preserving
933        the old behaviour.
934    
935    5.  The command line items --regex=pattern and --regexp=pattern were not
936        recognized by pcregrep, which required --regex pattern or --regexp pattern
937        (with a space rather than an '='). The man page documented the '=' forms,
938        which are compatible with GNU grep; these now work.
939    
940    6.  No libpcreposix.pc file was created for pkg-config; there was just
941        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
942    
943    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
944        when UCP support is not needed, by modifying the Python script that
945        generates it from Unicode data files. This should not matter if the module
946        is correctly used as a library, but I received one complaint about 50K of
947        unwanted data. My guess is that the person linked everything into his
948        program rather than using a library. Anyway, it does no harm.
949    
950    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
951        was a minimum greater than 1 for a wide character in a possessive
952        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
953        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
954        character. Chaos in the form of incorrect output or a compiling loop could
955        result.
956    
957    9.  The restrictions on what a pattern can contain when partial matching is
958        requested for pcre_exec() have been removed. All patterns can now be
959        partially matched by this function. In addition, if there are at least two
960        slots in the offset vector, the offset of the earliest inspected character
961        for the match and the offset of the end of the subject are set in them when
962        PCRE_ERROR_PARTIAL is returned.
963    
964    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
965        synonymous with PCRE_PARTIAL, for backwards compatibility, and
966        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
967        and may be more useful for multi-segment matching.
968    
969    11. Partial matching with pcre_exec() is now more intuitive. A partial match
970        used to be given if ever the end of the subject was reached; now it is
971        given only if matching could not proceed because another character was
972        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
973        string "Z", which now yields "no match" instead of "partial match". In the
974        case of pcre_dfa_exec(), "no match" is given if every matching path for the
975        final character ended with (*FAIL).
976    
977    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
978        if the pattern had a "must contain" character that was already found in the
979        earlier partial match, unless partial matching was again requested. For
980        example, with the pattern /dog.(body)?/, the "must contain" character is
981        "g". If the first part-match was for the string "dog", restarting with
982        "sbody" failed. This bug has been fixed.
983    
984    13. The string returned by pcre_dfa_exec() after a partial match has been
985        changed so that it starts at the first inspected character rather than the
986        first character of the match. This makes a difference only if the pattern
987        starts with a lookbehind assertion or \b or \B (\K is not supported by
988        pcre_dfa_exec()). It's an incompatible change, but it makes the two
989        matching functions compatible, and I think it's the right thing to do.
990    
991    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
992        so that the demonstration program is easily available in environments where
993        PCRE has not been installed from source.
994    
995    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
996        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
997        library.
998    
999    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1000        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1001        is not the first non-POSIX option to be added. Clearly some people find
1002        these options useful.
1003    
1004    17. If a caller to the POSIX matching function regexec() passes a non-zero
1005        value for nmatch with a NULL value for pmatch, the value of
1006        nmatch is forced to zero.
1007    
1008    18. RunGrepTest did not have a test for the availability of the -u option of
1009        the diff command, as RunTest does. It now checks in the same way as
1010        RunTest, and also checks for the -b option.
1011    
1012    19. If an odd number of negated classes containing just a single character
1013        interposed, within parentheses, between a forward reference to a named
1014        subpattern and the definition of the subpattern, compilation crashed with
1015        an internal error, complaining that it could not find the referenced
1016        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1017        [The bug was that it was starting one character too far in when skipping
1018        over the character class, thus treating the ] as data rather than
1019        terminating the class. This meant it could skip too much.]
1020    
1021    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1022        /g option in pcretest when the pattern contains \K, which makes it possible
1023        to have an empty string match not at the start, even when the pattern is
1024        anchored. Updated pcretest and pcredemo to use this option.
1025    
1026    21. If the maximum number of capturing subpatterns in a recursion was greater
1027        than the maximum at the outer level, the higher number was returned, but
1028        with unset values at the outer level. The correct (outer level) value is
1029        now given.
1030    
1031    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1032        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1033        make it do so. The string so far is captured, making this feature
1034        compatible with Perl.
1035    
1036    23. The tests have been re-organized, adding tests 11 and 12, to make it
1037        possible to check the Perl 5.10 features against Perl 5.10.
1038    
1039    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1040        pattern matches a fixed length string. PCRE did not allow this; now it
1041        does. Neither allows recursion.
1042    
1043    25. I finally figured out how to implement a request to provide the minimum
1044        length of subject string that was needed in order to match a given pattern.
1045        (It was back references and recursion that I had previously got hung up
1046        on.) This code has now been added to pcre_study(); it finds a lower bound
1047        to the length of subject needed. It is not necessarily the greatest lower
1048        bound, but using it to avoid searching strings that are too short does give
1049        some useful speed-ups. The value is available to calling programs via
1050        pcre_fullinfo().
1051    
1052    26. While implementing 25, I discovered to my embarrassment that pcretest had
1053        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1054        study optimizations had never been tested with that matching function.
1055        Oops. What is worse, even when it was passed study data, there was a bug in
1056        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1057        were also very few tests of studied patterns with pcre_dfa_exec().
1058    
1059    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1060        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1061        on the other side of the coin, they are no longer allowed to have different
1062        names, because these cannot be distinguished in PCRE, and this has caused
1063        confusion. (This is a difference from Perl.)
1064    
1065    28. When duplicate subpattern names are present (necessarily with different
1066        numbers, as required by 27 above), and a test is made by name in a
1067        conditional pattern, either for a subpattern having been matched, or for
1068        recursion in such a pattern, all the associated numbered subpatterns are
1069        tested, and the overall condition is true if the condition is true for any
1070        one of them. This is the way Perl works, and is also more like the way
1071        testing by number works.
1072    
1073    
1074    Version 7.9 11-Apr-09
1075    ---------------------
1076    
1077    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1078        (pcretest), all targets were linked against these libraries. This included
1079        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1080        libraries. This caused unwanted dependencies to be created. This problem
1081        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1082        pcretest is linked with readline.
1083    
1084    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1085        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1086        moved outside it again, because FALSE and TRUE are already defined in AIX,
1087        but BOOL is not.
1088    
1089    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1090        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1091    
1092    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1093        hyphens) following file names and line numbers when outputting matching
1094        lines. This is not true; no spaces are inserted. I have also clarified the
1095        wording for the --colour (or --color) option.
1096    
1097    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1098        was not coloured; this is different to GNU grep, so I have changed it to be
1099        the same.
1100    
1101    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1102        each matching line was coloured. Now it goes on to look for further matches
1103        of any of the test patterns, which is the same behaviour as GNU grep.
1104    
1105    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1106        doesn't make sense to accept an empty string match in pcregrep, so I have
1107        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1108        seems to be how GNU grep behaves.
1109    
1110    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1111        start or after a newline", because the conditional assertion was not being
1112        correctly handled. The rule now is that both the assertion and what follows
1113        in the first alternative must satisfy the test.
1114    
1115    9.  If auto-callout was enabled in a pattern with a conditional group whose
1116        condition was an assertion, PCRE could crash during matching, both with
1117        pcre_exec() and pcre_dfa_exec().
1118    
1119    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1120        used for matching.
1121    
1122    11. Unicode property support in character classes was not working for
1123        characters (bytes) greater than 127 when not in UTF-8 mode.
1124    
1125    12. Added the -M command line option to pcretest.
1126    
1127    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1128    
1129    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1130    
1131    16. Added comments and documentation about mis-use of no_arg in the C++
1132        wrapper.
1133    
1134    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1135        from Martin Jerabek that uses macro names for all relevant character and
1136        string constants.
1137    
1138    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1139        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1140        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1141        these, but not everybody uses configure.
1142    
1143    19. A conditional group that had only one branch was not being correctly
1144        recognized as an item that could match an empty string. This meant that an
1145        enclosing group might also not be so recognized, causing infinite looping
1146        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1147        with the subject "ab", where knowledge that the repeated group can match
1148        nothing is needed in order to break the loop.
1149    
1150    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1151        exec(), but without supplying a callout function, matching went wrong.
1152    
1153    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1154        leak if the size of the offset vector was greater than 30. When the vector
1155        is smaller, the saved offsets during recursion go onto a local stack
1156        vector, but for larger vectors malloc() is used. It was failing to free
1157        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1158        error, in fact).
1159    
1160    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1161        heapframe that is used only when UTF-8 support is enabled. This caused no
1162        problem, but was untidy.
1163    
1164    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1165        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1166        included within another project.
1167    
1168    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1169        slightly modified by me:
1170    
1171          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1172              not building pcregrep.
1173    
1174          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1175              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1176    
1177    25. Forward references, both numeric and by name, in patterns that made use of
1178        duplicate group numbers, could behave incorrectly or give incorrect errors,
1179        because when scanning forward to find the reference group, PCRE was not
1180        taking into account the duplicate group numbers. A pattern such as
1181        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1182    
1183    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1184        the feature of a custom pointer more persuasive (as requested by a user).
1185    
1186    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1187        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1188        pre-defined.
1189    
1190    28. Added support for (*UTF8) at the start of a pattern.
1191    
1192    29. Arrange for flags added by the "release type" setting in CMake to be shown
1193        in the configuration summary.
1194    
1195    
1196    Version 7.8 05-Sep-08
1197    ---------------------
1198    
1199    1.  Replaced UCP searching code with optimized version as implemented for Ad
1200        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1201        stage table and inline lookup instead of a function, giving speed ups of 2
1202        to 5 times on some simple patterns that I tested. Permission was given to
1203        distribute the MultiStage2.py script that generates the tables (it's not in
1204        the tarball, but is in the Subversion repository).
1205    
1206    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1207        scripts.
1208    
1209    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1210        a group with a zero qualifier. The result of the study could be incorrect,
1211        or the function might crash, depending on the pattern.
1212    
1213    4.  Caseless matching was not working for non-ASCII characters in back
1214        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1215        It now works when Unicode Property Support is available.
1216    
1217    5.  In pcretest, an escape such as \x{de} in the data was always generating
1218        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1219        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1220        truncation.
1221    
1222    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1223    
1224    7.  Added two (int) casts to pcregrep when printing the difference of two
1225        pointers, in case they are 64-bit values.
1226    
1227    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1228        test 2 if it fails.
1229    
1230    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1231        and a #define of that name to empty if it is not externally set. This is to
1232        allow users of MSVC to set it if necessary.
1233    
1234    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1235        the convenience functions in the pcre_get.c source file.
1236    
1237    11. An option change at the start of a pattern that had top-level alternatives
1238        could cause overwriting and/or a crash. This command provoked a crash in
1239        some environments:
1240    
1241          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1242    
1243        This potential security problem was recorded as CVE-2008-2371.
1244    
1245    12. For a pattern where the match had to start at the beginning or immediately
1246        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1247        pcre_dfa_exec() could read past the end of the passed subject if there was
1248        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1249        pcretest so that it places the subject at the end of its malloc-ed buffer.
1250    
1251    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1252        exec() might read past the end of the data buffer in UTF-8 mode.
1253    
1254    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1255        the data contained the byte 0x85 as part of a UTF-8 character within its
1256        first line. This applied both to normal and DFA matching.
1257    
1258    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1259        /^[^d]*?$/8 failed to match "abc".
1260    
1261    16. Added a missing copyright notice to pcrecpp_internal.h.
1262    
1263    17. Make it more clear in the documentation that values returned from
1264        pcre_exec() in ovector are byte offsets, not character counts.
1265    
1266    18. Tidied a few places to stop certain compilers from issuing warnings.
1267    
1268    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1269        supplied by Stefan Weber. I made a further small update for 7.8 because
1270        there is a change of source arrangements: the pcre_searchfuncs.c module is
1271        replaced by pcre_ucd.c.
1272    
1273    
1274    Version 7.7 07-May-08
1275    ---------------------
1276    
1277    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1278        a string to a long long, pretend we don't even have a long long." This is
1279        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1280    
1281    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1282        pre-7.6 versions, which defined a global no_arg variable instead of putting
1283        it in the RE class. (See also #8 below.)
1284    
1285    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1286        Lopes.
1287    
1288    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1289    
1290        (1) The include/exclude patterns were being applied to the whole pathnames
1291            of files, instead of just to the final components.
1292    
1293        (2) If there was more than one level of directory, the subdirectories were
1294            skipped unless they satisfied the include/exclude conditions. This is
1295            inconsistent with GNU grep (and could even be seen as contrary to the
1296            pcregrep specification - which I improved to make it absolutely clear).
1297            The action now is always to scan all levels of directory, and just
1298            apply the include/exclude patterns to regular files.
1299    
1300    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1301        --exclude_dir in the tests to avoid scanning .svn directories.
1302    
1303    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1304        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1305        doesn't support NULs in patterns.
1306    
1307    7.  Added some missing "const"s to declarations of static tables in
1308        pcre_compile.c and pcre_dfa_exec.c.
1309    
1310    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1311        caused by fix #2  above. (Subsequently also a second patch to fix the
1312        first patch. And a third patch - this was a messy problem.)
1313    
1314    9.  Applied Craig's patch to remove the use of push_back().
1315    
1316    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1317        matching function regexec().
1318    
1319    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1320        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1321        references. PCRE supports relative numbers with this syntax (I don't think
1322        Oniguruma does).
1323    
1324    12. Previously, a group with a zero repeat such as (...){0} was completely
1325        omitted from the compiled regex. However, this means that if the group
1326        was called as a subroutine from elsewhere in the pattern, things went wrong
1327        (an internal error was given). Such groups are now left in the compiled
1328        pattern, with a new opcode that causes them to be skipped at execution
1329        time.
1330    
1331    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1332        to the way PCRE behaves:
1333    
1334        (a) A lone ] character is dis-allowed (Perl treats it as data).
1335    
1336        (b) A back reference to an unmatched subpattern matches an empty string
1337            (Perl fails the current match path).
1338    
1339        (c) A data ] in a character class must be notated as \] because if the
1340            first data character in a class is ], it defines an empty class. (In
1341            Perl it is not possible to have an empty class.) The empty class []
1342            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1343            The negative empty class [^] matches any one character, independently
1344            of the DOTALL setting.
1345    
1346    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1347        non-existent subpattern following a character class starting with ']' and
1348        containing () gave an internal compiling error instead of "reference to
1349        non-existent subpattern". Fortunately, when the pattern did exist, the
1350        compiled code was correct. (When scanning forwards to check for the
1351        existencd of the subpattern, it was treating the data ']' as terminating
1352        the class, so got the count wrong. When actually compiling, the reference
1353        was subsequently set up correctly.)
1354    
1355    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1356        it was being rejected as not supported by pcre_dfa_exec(), even though
1357        other assertions are supported. I have made pcre_dfa_exec() support
1358        (*FAIL).
1359    
1360    16. The implementation of 13c above involved the invention of a new opcode,
1361        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1362        cannot be changed at match time, I realized I could make a small
1363        improvement to matching performance by compiling OP_ALLANY instead of
1364        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1365        on the OP_ANY path.
1366    
1367    17. Compiling pcretest on Windows with readline support failed without the
1368        following two fixes: (1) Make the unistd.h include conditional on
1369        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1370    
1371    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1372        ncurses library to be included for pcretest when ReadLine support is
1373        requested, but also to allow for it to be overridden. This patch came from
1374        Daniel Bergström.
1375    
1376    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1377        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1378        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1379        spotting this.
1380    
1381    
1382    Version 7.6 28-Jan-08
1383    ---------------------
1384    
1385    1.  A character class containing a very large number of characters with
1386        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1387        overflow.
1388    
1389    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1390        HAVE_LONG_LONG is not defined.
1391    
1392    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1393        bring them up to date and include new features. This patch includes:
1394    
1395        - Fixed PH's badly added libz and libbz2 support.
1396        - Fixed a problem with static linking.
1397        - Added pcredemo. [But later removed - see 7 below.]
1398        - Fixed dftables problem and added an option.
1399        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1400            HAVE_LONG_LONG.
1401        - Added readline support for pcretest.
1402        - Added an listing of the option settings after cmake has run.
1403    
1404    4.  A user submitted a patch to Makefile that makes it easy to create
1405        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1406        Makefile.am that cause it to include this special target, without
1407        affecting anything else. Note that the same mingw target plus all
1408        the other distribution libraries and programs are now supported
1409        when configuring with CMake (see 6 below) instead of with
1410        Configure/Make.
1411    
1412    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1413        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1414        exported in the Windows port". It has not yet been confirmed that the patch
1415        solves the problem, but it does no harm.
1416    
1417    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1418        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1419        with CMake, and also correct the comment about stack recursion.
1420    
1421    7.  Remove the automatic building of pcredemo from the ./configure system and
1422        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1423        of a program that users should build themselves after PCRE is installed, so
1424        building it automatically is not really right. What is more, it gave
1425        trouble in some build environments.
1426    
1427    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1428    
1429    
1430    Version 7.5 10-Jan-08
1431    ---------------------
1432    
1433    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1434        values in parens when parsing an RE using the C++ wrapper."
1435    
1436    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1437        Characters greater than 255 were excluded from the class instead of being
1438        included.
1439    
1440    3.  The same bug as (2) above applied to negated POSIX classes such as
1441        [:^space:].
1442    
1443    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1444        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1445        I have changed it.
1446    
1447    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1448        first named subpattern) and a construct such as (?&a) would reference the
1449        first named subpattern whose name started with "a" (in other words, the
1450        length check was missing). Both these problems are fixed. "Subpattern name
1451        expected" is now given for (?&) (a zero-length name), and this patch also
1452        makes it give the same error for \k'' (previously it complained that that
1453        was a reference to a non-existent subpattern).
1454    
1455    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1456        this is right because (?- can be followed by option settings as well as by
1457        digits. I have, however, made the messages clearer.
1458    
1459    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1460        than the number used in the conditional) now cause a compile-time error.
1461        This is actually not compatible with Perl, which accepts such patterns, but
1462        treats the conditional as always being FALSE (as PCRE used to), but it
1463        seems to me that giving a diagnostic is better.
1464    
1465    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1466        and messages.
1467    
1468    9.  Fix two occurrences of "backslash" in comments that should have been
1469        "backspace".
1470    
1471    10. Remove two redundant lines of code that can never be obeyed (their function
1472        was moved elsewhere).
1473    
1474    11. The program that makes PCRE's Unicode character property table had a bug
1475        which caused it to generate incorrect table entries for sequences of
1476        characters that have the same character type, but are in different scripts.
1477        It amalgamated them into a single range, with the script of the first of
1478        them. In other words, some characters were in the wrong script. There were
1479        thirteen such cases, affecting characters in the following ranges:
1480    
1481          U+002b0 - U+002c1
1482          U+0060c - U+0060d
1483          U+0061e - U+00612
1484          U+0064b - U+0065e
1485          U+0074d - U+0076d
1486          U+01800 - U+01805
1487          U+01d00 - U+01d77
1488          U+01d9b - U+01dbf
1489          U+0200b - U+0200f
1490          U+030fc - U+030fe
1491          U+03260 - U+0327f
1492          U+0fb46 - U+0fbb1
1493          U+10450 - U+1049d
1494    
1495    12. The -o option (show only the matching part of a line) for pcregrep was not
1496        compatible with GNU grep in that, if there was more than one match in a
1497        line, it showed only the first of them. It now behaves in the same way as
1498        GNU grep.
1499    
1500    13. If the -o and -v options were combined for pcregrep, it printed a blank
1501        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1502        does the same. The return code can be used to tell if there were any
1503        non-matching lines.
1504    
1505    14. Added --file-offsets and --line-offsets to pcregrep.
1506    
1507    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1508        infinitely looping recursion. The bug was that positive lookaheads were not
1509        being skipped when checking for a possible empty match (negative lookaheads
1510        and both kinds of lookbehind were skipped).
1511    
1512    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1513        inclusion of <windows.h> to before rather than after the definition of
1514        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1515    
1516    17. Specifying a possessive quantifier with a specific limit for a Unicode
1517        character property caused pcre_compile() to compile bad code, which led at
1518        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1519        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1520        caused the error; without that there was no problem.
1521    
1522    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1523    
1524    19. Added --enable-pcretest-libreadline.
1525    
1526    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1527        RE::GlobalReplace(). As a result, the number of replacements returned was
1528        double what it should be. I removed one of the increments, but Craig sent a
1529        later patch that removed the other one (the right fix) and added unit tests
1530        that check the return values (which was not done before).
1531    
1532    21. Several CMake things:
1533    
1534        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1535            the names libpcre and libpcreposix, not just pcre and pcreposix.
1536    
1537        (2) The above change means that pcretest and pcregrep are now correctly
1538            linked with the newly-built libraries, not previously installed ones.
1539    
1540        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1541    
1542    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1543        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1544        UTF-8 newline character). The key issue is that the pattern starts .*;
1545        this means that the match must be either at the beginning, or after a
1546        newline. The bug was in the code for advancing after a failed match and
1547        checking that the new position followed a newline. It was not taking
1548        account of UTF-8 characters correctly.
1549    
1550    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1551        character classes. PCRE was not treating the sequence [:...:] as a
1552        character class unless the ... were all letters. Perl, however, seems to
1553        allow any characters between [: and :], though of course it rejects as
1554        unknown any "names" that contain non-letters, because all the known class
1555        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1556        for example, whereas PCRE did not - it did not recognize a POSIX character
1557        class. This seemed a bit dangerous, so the code has been changed to be
1558        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1559        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1560        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1561        Perl does, and where it didn't before.
1562    
1563    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1564        Windows environments %n is disabled by default.
1565    
1566    
1567    Version 7.4 21-Sep-07
1568    ---------------------
1569    
1570    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1571        means that a class such as [\s] counted as "explicit reference to CR or
1572        LF". That isn't really right - the whole point of the change was to try to
1573        help when there was an actual mention of one of the two characters. So now
1574        the change happens only if \r or \n (or a literal CR or LF) character is
1575        encountered.
1576    
1577    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1578        of both had grown to the point where there were only 3 bits left.
1579        Fortunately, there was spare space in the data structure, and so I have
1580        moved the internal flags into a new 16-bit field to free up more option
1581        bits.
1582    
1583    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1584        but did not set the internal JCHANGED flag - either of these is enough to
1585        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1586        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1587        start sets both bits.
1588    
1589    4.  Added options (at build time, compile time, exec time) to change \R from
1590        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1591    
1592    5.  doc/pcresyntax.html was missing from the distribution.
1593    
1594    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1595        compatibility, even though it is no longer used.
1596    
1597    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1598        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1599        windows.h file is present (where different names are used). [This was
1600        reversed later after testing - see 16 below.]
1601    
1602    8.  Changed all #include <config.h> to #include "config.h". There were also
1603        some further <pcre.h> cases that I changed to "pcre.h".
1604    
1605    9.  When pcregrep was used with the --colour option, it missed the line ending
1606        sequence off the lines that it output.
1607    
1608    10. It was pointed out to me that arrays of string pointers cause lots of
1609        relocations when a shared library is dynamically loaded. A technique of
1610        using a single long string with a table of offsets can drastically reduce
1611        these. I have refactored PCRE in four places to do this. The result is
1612        dramatic:
1613    
1614          Originally:                          290
1615          After changing UCP table:            187
1616          After changing error message table:   43
1617          After changing table of "verbs"       36
1618          After changing table of Posix names   22
1619    
1620        Thanks to the folks working on Gregex for glib for this insight.
1621    
1622    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1623        unicode-properties was also set.
1624    
1625    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1626    
1627    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1628        checked only for CRLF.
1629    
1630    14. Added casts to pcretest.c to avoid compiler warnings.
1631    
1632    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1633    
1634    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1635        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1636        entirely. This removes changes made in 7 above.
1637    
1638    17. The CMake files have been updated, and there is now more information about
1639        building with CMake in the NON-UNIX-USE document.
1640    
1641    
1642    Version 7.3 28-Aug-07
1643    ---------------------
1644    
1645     1. In the rejigging of the build system that eventually resulted in 7.1, the
1646        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1647        brackets there is not right, since it causes compilers to look for an
1648        installed pcre.h, not the version that is in the source that is being
1649        compiled (which of course may be different). I have changed it back to:
1650    
1651          #include "pcre.h"
1652    
1653        I have a vague recollection that the change was concerned with compiling in
1654        different directories, but in the new build system, that is taken care of
1655        by the VPATH setting the Makefile.
1656    
1657     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1658        when the subject happened to end in the byte 0x85 (e.g. if the last
1659        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1660        characters but of course it shouldn't be taken as a newline when it is part
1661        of another character. The bug was that, for an unlimited repeat of . in
1662        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1663        characters when looking for a newline.
1664    
1665     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1666    
1667     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1668        in debug output.
1669    
1670     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1671        long printing in the pcrecpp unittest when running under MinGW.
1672    
1673     6. ESC_K was left out of the EBCDIC table.
1674    
1675     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1676        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1677        limit also applies to "virtual nesting" when a pattern is recursive, and in
1678        this case 1000 isn't so big. I have been able to remove this limit at the
1679        expense of backing off one optimization in certain circumstances. Normally,
1680        when pcre_exec() would call its internal match() function recursively and
1681        immediately return the result unconditionally, it uses a "tail recursion"
1682        feature to save stack. However, when a subpattern that can match an empty
1683        string has an unlimited repetition quantifier, it no longer makes this
1684        optimization. That gives it a stack frame in which to save the data for
1685        checking that an empty string has been matched. Previously this was taken
1686        from the 1000-entry workspace that had been reserved. So now there is no
1687        explicit limit, but more stack is used.
1688    
1689     8. Applied Daniel's patches to solve problems with the import/export magic
1690        syntax that is required for Windows, and which was going wrong for the
1691        pcreposix and pcrecpp parts of the library. These were overlooked when this
1692        problem was solved for the main library.
1693    
1694     9. There were some crude static tests to avoid integer overflow when computing
1695        the size of patterns that contain repeated groups with explicit upper
1696        limits. As the maximum quantifier is 65535, the maximum group length was
1697        set at 30,000 so that the product of these two numbers did not overflow a
1698        32-bit integer. However, it turns out that people want to use groups that
1699        are longer than 30,000 bytes (though not repeat them that many times).
1700        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1701        made it possible to implement the integer overflow checks in a much more
1702        dynamic way, which I have now done. The artificial limitation on group
1703        length has been removed - we now have only the limit on the total length of
1704        the compiled pattern, which depends on the LINK_SIZE setting.
1705    
1706    10. Fixed a bug in the documentation for get/copy named substring when
1707        duplicate names are permitted. If none of the named substrings are set, the
1708        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1709        empty string.
1710    
1711    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1712        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1713        because the ] is interpreted as the first data character and the
1714        terminating ] is not found. PCRE has been made compatible with Perl in this
1715        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1716        cause memory overwriting.
1717    
1718    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1719        string has been matched (to stop an infinite loop). It was not recognizing
1720        a conditional subpattern that could match an empty string if that
1721        subpattern was within another subpattern. For example, it looped when
1722        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1723        condition was not nested. This bug has been fixed.
1724    
1725    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1726        past the start of the subject in the presence of bytes with the top bit
1727        set, for example "\x8aBCD".
1728    
1729    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1730        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1731    
1732    14. Optimized (?!) to (*FAIL).
1733    
1734    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1735        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1736        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1737        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1738        does: it's just the validity check that is more restrictive.
1739    
1740    16. Inserted checks for integer overflows during escape sequence (backslash)
1741        processing, and also fixed erroneous offset values for syntax errors during
1742        backslash processing.
1743    
1744    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1745        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1746    
1747    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1748        caused an overrun.
1749    
1750    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1751        something other than just ASCII characters) inside a group that had an
1752        unlimited repeat caused a loop at compile time (while checking to see
1753        whether the group could match an empty string).
1754    
1755    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1756        [\P{Any}] did so. (Error in the code for printing property names.)
1757    
1758    21. An orphan \E inside a character class could cause a crash.
1759    
1760    22. A repeated capturing bracket such as (A)? could cause a wild memory
1761        reference during compilation.
1762    
1763    23. There are several functions in pcre_compile() that scan along a compiled
1764        expression for various reasons (e.g. to see if it's fixed length for look
1765        behind). There were bugs in these functions when a repeated \p or \P was
1766        present in the pattern. These operators have additional parameters compared
1767        with \d, etc, and these were not being taken into account when moving along
1768        the compiled data. Specifically:
1769    
1770        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1771            length.
1772    
1773        (b) An item such as \pL+ within a repeated group could cause crashes or
1774            loops.
1775    
1776        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1777            "reference to non-existent subpattern" error.
1778    
1779        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1780    
1781    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1782        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1783    
1784    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1785    
1786    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1787        character were causing crashes (broken optimization).
1788    
1789    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1790        \p or \P) caused a compile-time loop.
1791    
1792    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1793        break. For example, the unstudied pattern [\r\n]A does not match the string
1794        "\r\nA" because change 7.0/46 below moves the current point on by two
1795        characters after failing to match at the start. However, the pattern \nA
1796        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1797        the same is true. There doesn't seem any very clean way out of this, but
1798        what I have chosen to do makes the common cases work: PCRE now takes note
1799        of whether there can be an explicit match for \r or \n anywhere in the
1800        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1801        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1802        pattern has explicit CR or LF references.
1803    
1804    29. Added (*CR) etc for changing newline setting at start of pattern.
1805    
1806    
1807    Version 7.2 19-Jun-07
1808    ---------------------
1809    
1810     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1811        which is apparently normally available under Windows.
1812    
1813     2. Re-jig the pcregrep tests with different newline settings in an attempt
1814        to make them independent of the local environment's newline setting.
1815    
1816     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1817    
1818     4. Some of the "internals" tests were previously cut out when the link size
1819        was not 2, because the output contained actual offsets. The recent new
1820        "Z" feature of pcretest means that these can be cut out, making the tests
1821        usable with all link sizes.
1822    
1823     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1824        stack recursion. This gives a massive performance boost under BSD, but just
1825        a small improvement under Linux. However, it saves one field in the frame
1826        in all cases.
1827    
1828     6. Added more features from the forthcoming Perl 5.10:
1829    
1830        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1831            recursion call. It refers to the nth most recently opened parentheses.
1832    
1833        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1834            to be opened parentheses.
1835    
1836        (c) Conditions that refer to capturing parentheses can be specified
1837            relatively, for example, (?(-2)... or (?(+3)...
1838    
1839        (d) \K resets the start of the current match so that everything before
1840            is not part of it.
1841    
1842        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1843    
1844        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1845            reference syntax.
1846    
1847        (g) (?| introduces a group in which the numbering of parentheses in each
1848            alternative starts with the same number.
1849    
1850        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1851    
1852     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1853        PCRE_INFO_JCHANGED.
1854    
1855     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1856        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1857        for detecting groups that can match an empty string.
1858    
1859     9. A pattern with a very large number of alternatives (more than several
1860        hundred) was running out of internal workspace during the pre-compile
1861        phase, where pcre_compile() figures out how much memory will be needed. A
1862        bit of new cunning has reduced the workspace needed for groups with
1863        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1864        workspace instead of running out of the 4096 that are available.
1865    
1866    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1867    
1868    11. Applied patch from Google to remove an optimization that didn't quite work.
1869        The report of the bug said:
1870    
1871          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1872          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1873          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1874    
1875    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1876        it matched the wrong number of bytes.
1877    
1878    
1879    Version 7.1 24-Apr-07
1880    ---------------------
1881    
1882     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1883        that is more "standard", making use of automake and other Autotools. There
1884        is some re-arrangement of the files and adjustment of comments consequent
1885        on this.
1886    
1887     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1888        for recursive directory scanning broke on some systems because the files
1889        are not scanned in any specific order and on different systems the order
1890        was different. A call to "sort" has been inserted into RunGrepTest for the
1891        approprate test as a short-term fix. In the longer term there may be an
1892        alternative.
1893    
1894     3. I had an email from Eric Raymond about problems translating some of PCRE's
1895        man pages to HTML (despite the fact that I distribute HTML pages, some
1896        people do their own conversions for various reasons). The problems
1897        concerned the use of low-level troff macros .br and .in. I have therefore
1898        removed all such uses from the man pages (some were redundant, some could
1899        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1900        HTML has been updated to handle .nf/.fi and to complain if it encounters
1901        .br or .in.
1902    
1903     4. Updated comments in configure.ac that get placed in config.h.in and also
1904        arranged for config.h to be included in the distribution, with the name
1905        config.h.generic, for the benefit of those who have to compile without
1906        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1907    
1908     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1909        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1910        makevp.bat for the current PCRE, using the additional files
1911        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1912    
1913     6. A Windows user reported a minor discrepancy with test 2, which turned out
1914        to be caused by a trailing space on an input line that had got lost in his
1915        copy. The trailing space was an accident, so I've just removed it.
1916    
1917     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1918        that is needed.
1919    
1920     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1921        as "const" (a) because they are and (b) because it helps the PHP
1922        maintainers who have recently made a script to detect big data structures
1923        in the php code that should be moved to the .rodata section. I remembered
1924        to update Builducptable as well, so it won't revert if ucptable.h is ever
1925        re-created.
1926    
1927     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1928        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1929        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1930        support is not required. This saves 1.5-2K of code, which is important in
1931        some applications.
1932    
1933        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1934        so as not to refer to the tables, even though these functions will never be
1935        called when UTF-8 support is disabled. Otherwise there are problems with a
1936        shared library.
1937    
1938    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1939    
1940        (a) It was defining its arguments as char * instead of void *.
1941    
1942        (b) It was assuming that all moves were upwards in memory; this was true
1943            a long time ago when I wrote it, but is no longer the case.
1944    
1945        The emulated memove() is provided for those environments that have neither
1946        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1947        is clearly not the case, as these two bugs were recently reported.
1948    
1949    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1950        and Detrail to create the HTML documentation, the .txt form of the man
1951        pages, and it removes trailing spaces from listed files. It also creates
1952        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1953        case, it wraps all the #defines with #ifndefs. This script should be run
1954        before "make dist".
1955    
1956    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1957        with Unicode property support.
1958    
1959        (a) For a maximizing quantifier, if the two different cases of the
1960            character were of different lengths in their UTF-8 codings (there are
1961            some cases like this - I found 11), and the matching function had to
1962            back up over a mixture of the two cases, it incorrectly assumed they
1963            were both the same length.
1964    
1965        (b) When PCRE was configured to use the heap rather than the stack for
1966            recursion during matching, it was not correctly preserving the data for
1967            the other case of a UTF-8 character when checking ahead for a match
1968            while processing a minimizing repeat. If the check also involved
1969            matching a wide character, but failed, corruption could cause an
1970            erroneous result when trying to check for a repeat of the original
1971            character.
1972    
1973    13. Some tidying changes to the testing mechanism:
1974    
1975        (a) The RunTest script now detects the internal link size and whether there
1976            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1977            values substituted by "configure". (The RunGrepTest script already did
1978            this for UTF-8.) The configure.ac script no longer substitutes the
1979            relevant variables.
1980    
1981        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1982            with length and offset values. This means that the output is different
1983            for different internal link sizes. Test 2 is skipped for link sizes
1984            other than 2 because of this, bypassing the problem. Unfortunately,
1985            there was also a test in test 3 (the locale tests) that used /B and
1986            failed for link sizes other than 2. Rather than cut the whole test out,
1987            I have added a new /Z option to pcretest that replaces the length and
1988            offset values with spaces. This is now used to make test 3 independent
1989            of link size. (Test 2 will be tidied up later.)
1990    
1991    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1992        segmentation fault instead of returning the appropriate error message.
1993    
1994    15. In multiline mode when the newline sequence was set to "any", the pattern
1995        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1996        This doesn't seem right; it now treats the CRLF combination as the line
1997        ending, and so does not match in that case. It's only a pattern such as ^$
1998        that would hit this one: something like ^ABC$ would have failed after \r
1999        and then tried again after \r\n.
2000    
2001    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2002        in an attempt to make files that differ only in their line terminators
2003        compare equal. This works on Linux.
2004    
2005    17. Under certain error circumstances pcregrep might try to free random memory
2006        as it exited. This is now fixed, thanks to valgrind.
2007    
2008    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2009        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2010        was because its rules for how to advance for /g after matching an empty
2011        string at the end of a line did not allow for this case. They now check for
2012        it specially.
2013    
2014    20. pcretest is supposed to handle patterns and data of any length, by
2015        extending its buffers when necessary. It was getting this wrong when the
2016        buffer for a data line had to be extended.
2017    
2018    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2019        CRLF as a newline sequence.
2020    
2021    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2022        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2023        I have nevertheless tidied it up.
2024    
2025    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2026    
2027    24. Added a man page for pcre-config.
2028    
2029    
2030    Version 7.0 19-Dec-06
2031    ---------------------
2032    
2033     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2034        moving to gcc 4.1.1.
2035    
2036     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2037        sys/time.h, which is documented as needed for this function. It doesn't
2038        seem to matter on Linux, but it showed up on some releases of OS X.
2039    
2040     3. It seems that there are systems where bytes whose values are greater than
2041        127 match isprint() in the "C" locale. The "C" locale should be the
2042        default when a C program starts up. In most systems, only ASCII printing
2043        characters match isprint(). This difference caused the output from pcretest
2044        to vary, making some of the tests fail. I have changed pcretest so that:
2045    
2046        (a) When it is outputting text in the compiled version of a pattern, bytes
2047            other than 32-126 are always shown as hex escapes.
2048    
2049        (b) When it is outputting text that is a matched part of a subject string,
2050            it does the same, unless a different locale has been set for the match
2051            (using the /L modifier). In this case, it uses isprint() to decide.
2052    
2053     4. Fixed a major bug that caused incorrect computation of the amount of memory
2054        required for a compiled pattern when options that changed within the
2055        pattern affected the logic of the preliminary scan that determines the
2056        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2057        that the computed length was too small. The symptoms of this bug were
2058        either the PCRE error "internal error: code overflow" from pcre_compile(),
2059        or a glibc crash with a message such as "pcretest: free(): invalid next
2060        size (fast)". Examples of patterns that provoked this bug (shown in
2061        pcretest format) are:
2062    
2063          /(?-x: )/x
2064          /(?x)(?-x: \s*#\s*)/
2065          /((?i)[\x{c0}])/8
2066          /(?i:[\x{c0}])/8
2067    
2068        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2069        is now done differently.
2070    
2071     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2072        wrapper classes; (b) implement a new function in the C++ scanner that is
2073        more efficient than the old way of doing things because it avoids levels of
2074        recursion in the regex matching; (c) add a paragraph to the documentation
2075        for the FullMatch() function.
2076    
2077     6. The escape sequence \n was being treated as whatever was defined as
2078        "newline". Not only was this contrary to the documentation, which states
2079        that \n is character 10 (hex 0A), but it also went horribly wrong when
2080        "newline" was defined as CRLF. This has been fixed.
2081    
2082     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2083        was being set to -1 for the "end of line" case (supposedly a value that no
2084        character can have). Though this value is never used (the check for end of
2085        line is "zero bytes in current character"), it caused compiler complaints.
2086        I've changed it to 0xffffffff.
2087    
2088     8. In pcre_version.c, the version string was being built by a sequence of
2089        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2090        string (as it is for production releases) called a macro with an empty
2091        argument. The C standard says the result of this is undefined. The gcc
2092        compiler treats it as an empty string (which was what was wanted) but it is
2093        reported that Visual C gives an error. The source has been hacked around to
2094        avoid this problem.
2095    
2096     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2097        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2098        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2099        of them did).
2100    
2101    10. Originally, pcretest opened its input and output without "b"; then I was
2102        told that "b" was needed in some environments, so it was added for release
2103        5.0 to both the input and output. (It makes no difference on Unix-like
2104        systems.) Later I was told that it is wrong for the input on Windows. I've
2105        now abstracted the modes into two macros, to make it easier to fiddle with
2106        them, and removed "b" from the input mode under Windows.
2107    
2108    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2109    
2110    12. Added -help and --help to pcretest as an official way of being reminded
2111        of the options.
2112    
2113    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2114        and pcrecpp.cc because they annoy compilers at high warning levels.
2115    
2116    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2117    
2118    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2119        scripts are not C programs :-) and which was not noticed because it works
2120        on Linux.
2121    
2122    16. pcretest is supposed to handle any length of pattern and data line (as one
2123        line or as a continued sequence of lines) by extending its input buffer if
2124        necessary. This feature was broken for very long pattern lines, leading to
2125        a string of junk being passed to pcre_compile() if the pattern was longer
2126        than about 50K.
2127    
2128    17. I have done a major re-factoring of the way pcre_compile() computes the
2129        amount of memory needed for a compiled pattern. Previously, there was code
2130        that made a preliminary scan of the pattern in order to do this. That was
2131        OK when PCRE was new, but as the facilities have expanded, it has become
2132        harder and harder to keep it in step with the real compile phase, and there
2133        have been a number of bugs (see for example, 4 above). I have now found a
2134        cunning way of running the real compile function in a "fake" mode that
2135        enables it to compute how much memory it would need, while actually only
2136        ever using a few hundred bytes of working memory and without too many
2137        tests of the mode. This should make future maintenance and development
2138        easier. A side effect of this work is that the limit of 200 on the nesting
2139        depth of parentheses has been removed (though this was never a serious
2140        limitation, I suspect). However, there is a downside: pcre_compile() now
2141        runs more slowly than before (30% or more, depending on the pattern). I
2142        hope this isn't a big issue. There is no effect on runtime performance.
2143    
2144    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2145        newline (only possible for the last line of a file) and it was a
2146        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2147    
2148    19. Added additional timing features to pcretest. (1) The -tm option now times
2149        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2150        separate command line item, by a number that specifies the number of
2151        repeats to use when timing. The default is 50000; this gives better
2152        precision, but takes uncomfortably long for very large patterns.
2153    
2154    20. Extended pcre_study() to be more clever in cases where a branch of a
2155        subpattern has no definite first character. For example, (a*|b*)[cd] would
2156        previously give no result from pcre_study(). Now it recognizes that the
2157        first character must be a, b, c, or d.
2158    
2159    21. There was an incorrect error "recursive call could loop indefinitely" if
2160        a subpattern (or the entire pattern) that was being tested for matching an
2161        empty string contained only one non-empty item after a nested subpattern.
2162        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2163        incorrectly, because the \d was being skipped in the check.
2164    
2165    22. The pcretest program now has a new pattern option /B and a command line
2166        option -b, which is equivalent to adding /B to every pattern. This causes
2167        it to show the compiled bytecode, without the additional information that
2168        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2169        is the same as /B/I).
2170    
2171    23. A new optimization is now able automatically to treat some sequences such
2172        as a*b as a*+b. More specifically, if something simple (such as a character
2173        or a simple class like \d) has an unlimited quantifier, and is followed by
2174        something that cannot possibly match the quantified thing, the quantifier
2175        is automatically "possessified".
2176    
2177    24. A recursive reference to a subpattern whose number was greater than 39
2178        went wrong under certain circumstances in UTF-8 mode. This bug could also
2179        have affected the operation of pcre_study().
2180    
2181    25. Realized that a little bit of performance could be had by replacing
2182        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2183    
2184    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2185    
2186    27. Possessive quantifiers such as a++ were previously implemented by turning
2187        them into atomic groups such as ($>a+). Now they have their own opcodes,
2188        which improves performance. This includes the automatically created ones
2189        from 23 above.
2190    
2191    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2192        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2193        the first matched character to be a colon. This applied both to named and
2194        numbered groups.
2195    
2196    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2197    
2198    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2199        building PCRE on Windows easier, so I have included it in the distribution.
2200    
2201    31. There is now a check in pcretest against a ridiculously large number being
2202        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2203        loop, the loop is abandoned.
2204    
2205    32. Forward references to subpatterns in conditions such as (?(2)...) where
2206        subpattern 2 is defined later cause pcre_compile() to search forwards in
2207        the pattern for the relevant set of parentheses. This search went wrong
2208        when there were unescaped parentheses in a character class, parentheses
2209        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2210    
2211    33. "Subroutine" calls and backreferences were previously restricted to
2212        referencing subpatterns earlier in the regex. This restriction has now
2213        been removed.
2214    
2215    34. Added a number of extra features that are going to be in Perl 5.10. On the
2216        whole, these are just syntactic alternatives for features that PCRE had
2217        previously implemented using the Python syntax or my own invention. The
2218        other formats are all retained for compatibility.
2219    
2220        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2221            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2222            also .NET compatible.
2223    
2224        (b) A recursion or subroutine call to a named group can now be defined as
2225            (?&name) as well as (?P>name).
2226    
2227        (c) A backreference to a named group can now be defined as \k<name> or
2228            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2229            5.10, are also .NET compatible.
2230    
2231        (d) A conditional reference to a named group can now use the syntax
2232            (?(<name>) or (?('name') as well as (?(name).
2233    
2234        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2235            groups (named and numbered) that are never evaluated inline, but can be
2236            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2237            is always false. There may be only one alternative in such a group.
2238    
2239        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2240            as the simple (?(R). The condition is true only if the most recent
2241            recursion is that of the given number or name. It does not search out
2242            through the entire recursion stack.
2243    
2244        (g) The escape \gN or \g{N} has been added, where N is a positive or
2245            negative number, specifying an absolute or relative reference.
2246    
2247    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2248        some "unreachable code" warnings.
2249    
2250    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2251        things, this adds five new scripts.
2252    
2253    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2254        There were also incompatibilities regarding the handling of \Q..\E inside
2255        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2256        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2257    
2258    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2259        matches an empty string, and forcibly breaks the loop. There were bugs in
2260        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2261        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2262        separate and independent bugs (that affected different cases) have been
2263        fixed.
2264    
2265    39. Refactored the code to abolish the use of different opcodes for small
2266        capturing bracket numbers. This is a tidy that I avoided doing when I
2267        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2268        The new approach is not only tidier, it makes it possible to reduce the
2269        memory needed to fix the previous bug (38).
2270    
2271    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2272        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2273        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2274        mode.
2275    
2276    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2277        report.
2278    
2279    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2280        copy construction and assignment in the C++ wrapper.
2281    
2282    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2283        couple of bugs that could have given wrong results in the "--newline=crlf"
2284        case.
2285    
2286    44. Added a number of casts and did some reorganization of signed/unsigned int
2287        variables following suggestions from Dair Grant. Also renamed the variable
2288        "this" as "item" because it is a C++ keyword.
2289    
2290    45. Arranged for dftables to add
2291    
2292          #include "pcre_internal.h"
2293    
2294        to pcre_chartables.c because without it, gcc 4.x may remove the array
2295        definition from the final binary if PCRE is built into a static library and
2296        dead code stripping is activated.
2297    
2298    46. For an unanchored pattern, if a match attempt fails at the start of a
2299        newline sequence, and the newline setting is CRLF or ANY, and the next two
2300        characters are CRLF, advance by two characters instead of one.
2301    
2302    
2303    Version 6.7 04-Jul-06
2304    ---------------------
2305    
2306     1. In order to handle tests when input lines are enormously long, pcretest has
2307        been re-factored so that it automatically extends its buffers when
2308        necessary. The code is crude, but this _is_ just a test program. The
2309        default size has been increased from 32K to 50K.
2310    
2311     2. The code in pcre_study() was using the value of the re argument before
2312        testing it for NULL. (Of course, in any sensible call of the function, it
2313        won't be NULL.)
2314    
2315     3. The memmove() emulation function in pcre_internal.h, which is used on
2316        systems that lack both memmove() and bcopy() - that is, hardly ever -
2317        was missing a "static" storage class specifier.
2318    
2319     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2320        containing an extended class (one that cannot be represented by a bitmap
2321        because it contains high-valued characters or Unicode property items, e.g.
2322        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2323        pattern, but PCRE should not loop if you do not (it no longer does).
2324        [Detail: two cases were found: (a) a repeated subpattern containing an
2325        extended class; (b) a recursive reference to a subpattern that followed a
2326        previous extended class. It wasn't skipping over the extended class
2327        correctly when UTF-8 mode was not set.]
2328    
2329     5. A negated single-character class was not being recognized as fixed-length
2330        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2331        compile error "lookbehind assertion is not fixed length".
2332    
2333     6. The RunPerlTest auxiliary script was showing an unexpected difference
2334        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2335        write a Perl script that can interpret lines of an input file either as
2336        byte characters or as UTF-8, which is what "perltest" was being required to
2337        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2338        can't do is switch easily at run time between having the "use utf8;" pragma
2339        or not. In the end, I fudged it by using the RunPerlTest script to insert
2340        "use utf8;" explicitly for the UTF-8 tests.
2341    
2342     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2343        the end of the subject string, contrary to the documentation and to what
2344        Perl does. This was true of both matching functions. Now it matches only at
2345        the start of the subject and immediately after *internal* newlines.
2346    
2347     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2348        a pointer to an int instead of a pointer to an unsigned long int. This
2349        caused problems on 64-bit systems.
2350    
2351     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2352        instance of the 'standard' template library not being so standard".
2353    
2354    10. There was no check on the number of named subpatterns nor the maximum
2355        length of a subpattern name. The product of these values is used to compute
2356        the size of the memory block for a compiled pattern. By supplying a very
2357        long subpattern name and a large number of named subpatterns, the size
2358        computation could be caused to overflow. This is now prevented by limiting
2359        the length of names to 32 characters, and the number of named subpatterns
2360        to 10,000.
2361    
2362    11. Subpatterns that are repeated with specific counts have to be replicated in
2363        the compiled pattern. The size of memory for this was computed from the
2364        length of the subpattern and the repeat count. The latter is limited to
2365        65535, but there was no limit on the former, meaning that integer overflow
2366        could in principle occur. The compiled length of a repeated subpattern is
2367        now limited to 30,000 bytes in order to prevent this.
2368    
2369    12. Added the optional facility to have named substrings with the same name.
2370    
2371    13. Added the ability to use a named substring as a condition, using the
2372        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2373        are numbers (not recommended). Forward references are permitted.
2374    
2375    14. Added forward references in named backreferences (if you see what I mean).
2376    
2377    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2378        pattern could run off the end of the subject. For example, the pattern
2379        "(?s)(.{1,5})"8 did this with the subject "ab".
2380    
2381    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2382        PCRE_CASELESS was set when matching characters that were quantified with ?
2383        or *.
2384    
2385    17. A character class other than a single negated character that had a minimum
2386        but no maximum quantifier - for example [ab]{6,} - was not handled
2387        correctly by pce_dfa_exec(). It would match only one character.
2388    
2389    18. A valid (though odd) pattern that looked like a POSIX character
2390        class but used an invalid character after [ (for example [[,abc,]]) caused
2391        pcre_compile() to give the error "Failed: internal error: code overflow" or
2392        in some cases to crash with a glibc free() error. This could even happen if
2393        the pattern terminated after [[ but there just happened to be a sequence of
2394        letters, a binary zero, and a closing ] in the memory that followed.
2395    
2396    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2397        over the years. Originally (before any Unicode support), just the bottom 8
2398        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2399        output from "man perlunicode" includes this:
2400    
2401          The regular expression compiler produces polymorphic opcodes.  That
2402          is, the pattern adapts to the data and automatically switches to
2403          the Unicode character scheme when presented with Unicode data--or
2404          instead uses a traditional byte scheme when presented with byte
2405          data.
2406    
2407        Sadly, a wide octal escape does not cause a switch, and in a string with
2408        no other multibyte characters, these octal escapes are treated as before.
2409        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2410        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2411        Unicode string.
2412    
2413        I have not perpetrated such confusion in PCRE. Up till now, it took just
2414        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2415        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2416        translate to the appropriate multibyte character.
2417    
2418    29. Applied some refactoring to reduce the number of warnings from Microsoft
2419        and Borland compilers. This has included removing the fudge introduced
2420        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2421        a warning about an unused variable.
2422    
2423    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2424        characters since release 4.0, because Perl (from release 5.004) does not.
2425        [Or at least, is documented not to: some releases seem to be in conflict
2426        with the documentation.] However, when a pattern was studied with
2427        pcre_study() and all its branches started with \s, PCRE still included VT
2428        as a possible starting character. Of course, this did no harm; it just
2429        caused an unnecessary match attempt.
2430    
2431    22. Removed a now-redundant internal flag bit that recorded the fact that case
2432        dependency changed within the pattern. This was once needed for "required
2433        byte" processing, but is no longer used. This recovers a now-scarce options
2434        bit. Also moved the least significant internal flag bit to the most-
2435        significant bit of the word, which was not previously used (hangover from
2436        the days when it was an int rather than a uint) to free up another bit for
2437        the future.
2438    
2439    23. Added support for CRLF line endings as well as CR and LF. As well as the
2440        default being selectable at build time, it can now be changed at runtime
2441        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2442        specify that it is scanning data with non-default line endings.
2443    
2444    24. Changed the definition of CXXLINK to make it agree with the definition of
2445        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2446    
2447    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2448        recursions. This makes a big different to stack usage for some patterns.
2449    
2450    26. If a subpattern containing a named recursion or subroutine reference such
2451        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2452        the space required for the compiled pattern went wrong and gave too small a
2453        value. Depending on the environment, this could lead to "Failed: internal
2454        error: code overflow at offset 49" or "glibc detected double free or
2455        corruption" errors.
2456    
2457    27. Applied patches from Google (a) to support the new newline modes and (b) to
2458        advance over multibyte UTF-8 characters in GlobalReplace.
2459    
2460    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2461        difference for some implementation of PCRE in some Windows version.
2462    
2463    29. Added some extra testing facilities to pcretest:
2464    
2465        \q<number>   in a data line sets the "match limit" value
2466        \Q<number>   in a data line sets the "match recursion limt" value
2467        -S <number>  sets the stack size, where <number> is in megabytes
2468    
2469        The -S option isn't available for Windows.
2470    
2471    
2472    Version 6.6 06-Feb-06
2473    ---------------------
2474    
2475     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2476        in pcreposix.h. I have copied the definition from pcre.h.
2477    
2478     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2479        because pcre.h is no longer a built file.
2480    
2481     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2482        not normally included in the compiled code.
2483    
2484    
2485    Version 6.5 01-Feb-06
2486    ---------------------
2487    
2488     1. When using the partial match feature with pcre_dfa_exec(), it was not
2489        anchoring the second and subsequent partial matches at the new starting
2490        point. This could lead to incorrect results. For example, with the pattern
2491        /1234/, partially matching against "123" and then "a4" gave a match.
2492    
2493     2. Changes to pcregrep:
2494    
2495        (a) All non-match returns from pcre_exec() were being treated as failures
2496            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2497            error message is output. Some extra information is given for the
2498            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2499            probably the only errors that are likely to be caused by users (by
2500            specifying a regex that has nested indefinite repeats, for instance).
2501            If there are more than 20 of these errors, pcregrep is abandoned.
2502    
2503        (b) A binary zero was treated as data while matching, but terminated the
2504            output line if it was written out. This has been fixed: binary zeroes
2505            are now no different to any other data bytes.
2506    
2507        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2508            used to set a locale for matching. The --locale=xxxx long option has
2509            been added (no short equivalent) to specify a locale explicitly on the
2510            pcregrep command, overriding the environment variables.
2511    
2512        (d) When -B was used with -n, some line numbers in the output were one less
2513            than they should have been.
2514    
2515        (e) Added the -o (--only-matching) option.
2516    
2517        (f) If -A or -C was used with -c (count only), some lines of context were
2518            accidentally printed for the final match.
2519    
2520        (g) Added the -H (--with-filename) option.
2521    
2522        (h) The combination of options -rh failed to suppress file names for files
2523            that were found from directory arguments.
2524    
2525        (i) Added the -D (--devices) and -d (--directories) options.
2526    
2527        (j) Added the -F (--fixed-strings) option.
2528    
2529        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2530    
2531        (l) Added the --colo(u)r option.
2532    
2533        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2534            is not present by default.
2535    
2536     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2537        items such as (?R) or (?1), when the recursion could match a number of
2538        alternatives. If it matched one of the alternatives, but subsequently,
2539        outside the recursion, there was a failure, the code tried to back up into
2540        the recursion. However, because of the way PCRE is implemented, this is not
2541        possible, and the result was an incorrect result from the match.
2542    
2543        In order to prevent this happening, the specification of recursion has
2544        been changed so that all such subpatterns are automatically treated as
2545        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2546    
2547     4. I had overlooked the fact that, in some locales, there are characters for
2548        which isalpha() is true but neither isupper() nor islower() are true. In
2549        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2550        and ordfeminine) are like this. This affected the treatment of \w and \W
2551        when they appeared in character classes, but not when they appeared outside
2552        a character class. The bit map for "word" characters is now created
2553        separately from the results of isalnum() instead of just taking it from the
2554        upper, lower, and digit maps. (Plus the underscore character, of course.)
2555    
2556     5. The above bug also affected the handling of POSIX character classes such as
2557        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2558        permanent tables. Instead, the bit maps for such a class were previously
2559        created as the appropriate unions of the upper, lower, and digit bitmaps.
2560        Now they are created by subtraction from the [[:word:]] class, which has
2561        its own bitmap.
2562    
2563     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2564        It is created by subtracting the vertical space characters (\x09, \x0a,
2565        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2566        subtraction was done in the overall bitmap for a character class, meaning
2567        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2568        be recognized. This bug has been fixed.
2569    
2570     7. Patches from the folks at Google:
2571    
2572          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2573          real life, but is still worth protecting against".
2574    
2575          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2576          regular expressions".
2577    
2578          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2579          have it.
2580    
2581          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2582          "configure" and the latter not, in order to fix a problem somebody had
2583          with compiling the Arg class on HP-UX.
2584    
2585          (e) Improve the error-handling of the C++ wrapper a little bit.
2586    
2587          (f) New tests for checking recursion limiting.
2588    
2589     8. The pcre_memmove() function, which is used only if the environment does not
2590        have a standard memmove() function (and is therefore rarely compiled),
2591        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2592        returning a result (though PCRE never actually uses the result).
2593    
2594     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2595        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2596        returned instead of calling malloc() with an overflowing number that would
2597        most likely cause subsequent chaos.
2598    
2599    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2600    
2601    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2602        with this option is matched, the nmatch and pmatch options of regexec() are
2603        ignored.
2604    
2605    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2606        provided in case anyone wants to the the POSIX interface with UTF-8
2607        strings.
2608    
2609    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2610        C++ linking (needed for some HP-UX environments).
2611    
2612    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2613        (unused parameter) and in the pcre_printint() function (omitted "default"
2614        switch label when the default is to do nothing).
2615    
2616    15. Added some code to make it possible, when PCRE is compiled as a C++
2617        library, to replace subject pointers for pcre_exec() with a smart pointer
2618        class, thus making it possible to process discontinuous strings.
2619    
2620    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2621        much the same function. They were added by different people who were trying
2622        to make PCRE easy to compile on non-Unix systems. It has been suggested
2623        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2624        for compiling on Windows systems. I have therefore replaced it with
2625        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2626        defaults to "extern" for C or "extern C" for C++, which works fine on
2627        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2628        SCOPE with something explicit in config.h. In addition:
2629    
2630        (a) pcreposix.h still had just "extern" instead of either of these macros;
2631            I have replaced it with PCRE_DATA_SCOPE.
2632    
2633        (b) Functions such as _pcre_xclass(), which are internal to the library,
2634            but external in the C sense, all had PCRE_EXPORT in their definitions.
2635            This is apparently wrong for the Windows case, so I have removed it.
2636            (It makes no difference on Unix-like systems.)
2637    
2638    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2639        of recursive calls to match(). This is different to MATCH_LIMIT because
2640        that limits the total number of calls to match(), not all of which increase
2641        the depth of recursion. Limiting the recursion depth limits the amount of
2642        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2643        when PCRE is compiled, and changed at run time. A patch from Google adds
2644        this functionality to the C++ interface.
2645    
2646    18. Changes to the handling of Unicode character properties:
2647    
2648        (a) Updated the table to Unicode 4.1.0.
2649    
2650        (b) Recognize characters that are not in the table as "Cn" (undefined).
2651    
2652        (c) I revised the way the table is implemented to a much improved format
2653            which includes recognition of ranges. It now supports the ranges that
2654            are defined in UnicodeData.txt, and it also amalgamates other
2655            characters into ranges. This has reduced the number of entries in the
2656            table from around 16,000 to around 3,000, thus reducing its size
2657            considerably. I realized I did not need to use a tree structure after
2658            all - a binary chop search is just as efficient. Having reduced the
2659            number of entries, I extended their size from 6 bytes to 8 bytes to
2660            allow for more data.
2661    
2662        (d) Added support for Unicode script names via properties such as \p{Han}.
2663    
2664    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2665        matching that character.
2666    
2667    20. When matching a repeated Unicode property with a minimum greater than zero,
2668        (for example \pL{2,}), PCRE could look past the end of the subject if it
2669        reached it while seeking the minimum number of characters. This could
2670        happen only if some of the characters were more than one byte long, because
2671        there is a check for at least the minimum number of bytes.
2672    
2673    21. Refactored the implementation of \p and \P so as to be more general, to
2674        allow for more different types of property in future. This has changed the
2675        compiled form incompatibly. Anybody with saved compiled patterns that use
2676        \p or \P will have to recompile them.
2677    
2678    22. Added "Any" and "L&" to the supported property types.
2679    
2680    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2681        but give a compile time error if the value is greater than 0xff.
2682    
2683    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2684        accidentally not being installed or uninstalled.
2685    
2686    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2687        made were to insert the current release number. This seemed silly, because
2688        it made things harder for people building PCRE on systems that don't run
2689        "configure". I have turned pcre.h into a distributed file, no longer built
2690        by "configure", with the version identification directly included. There is
2691        no longer a pcre.h.in file.
2692    
2693        However, this change necessitated a change to the pcre-config script as
2694        well. It is built from pcre-config.in, and one of the substitutions was the
2695        release number. I have updated configure.ac so that ./configure now finds
2696        the release number by grepping pcre.h.
2697    
2698    26. Added the ability to run the tests under valgrind.
2699    
2700    
2701    Version 6.4 05-Sep-05
2702    ---------------------
2703    
2704     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2705        "--" to be printed when multiple files were scanned, even when none of the
2706        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2707        consider it to be a bug, and have restored the previous behaviour.
2708    
2709     2. A couple of code tidies to get rid of compiler warnings.
2710    
2711     3. The pcretest program used to cheat by referring to symbols in the library
2712        whose names begin with _pcre_. These are internal symbols that are not
2713        really supposed to be visible externally, and in some environments it is
2714        possible to suppress them. The cheating is now confined to including
2715        certain files from the library's source, which is a bit cleaner.
2716    
2717     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2718        file's purpose clearer.
2719    
2720     5. Reorganized pcre_ucp_findchar().
2721    
2722    
2723    Version 6.3 15-Aug-05
2724    ---------------------
2725    
2726     1. The file libpcre.pc.in did not have general read permission in the tarball.
2727    
2728     2. There were some problems when building without C++ support:
2729    
2730        (a) If C++ support was not built, "make install" and "make test" still
2731            tried to test it.
2732    
2733        (b) There were problems when the value of CXX was explicitly set. Some
2734            changes have been made to try to fix these, and ...
2735    
2736        (c) --disable-cpp can now be used to explicitly disable C++ support.
2737    
2738        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2739            backslash in a target when C++ was disabled. This confuses some
2740            versions of "make", apparently. Using an intermediate variable solves
2741            this. (Same for CPP_LOBJ.)
2742    
2743     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2744        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2745        necessary on certain architectures.
2746    
2747     4. Added a setting of -export-symbols-regex to the link command to remove
2748        those symbols that are exported in the C sense, but actually are local
2749        within the library, and not documented. Their names all begin with
2750        "_pcre_". This is not a perfect job, because (a) we have to except some
2751        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2752        available (and never for static libraries). I have made a note to try to
2753        find a way round (a) in the future.
2754    
2755    
2756    Version 6.2 01-Aug-05
2757    ---------------------
2758    
2759     1. There was no test for integer overflow of quantifier values. A construction
2760        such as {1111111111111111} would give undefined results. What is worse, if
2761        a minimum quantifier for a parenthesized subpattern overflowed and became
2762        negative, the calculation of the memory size went wrong. This could have
2763        led to memory overwriting.
2764    
2765     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2766    
2767     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2768        operating environments where this matters.
2769    
2770     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2771        PCRE options from within the C++ wrapper.
2772    
2773     5. Named capturing subpatterns were not being correctly counted when a pattern
2774        was compiled. This caused two problems: (a) If there were more than 100
2775        such subpatterns, the calculation of the memory needed for the whole
2776        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2777        back references of the form \12, where the number was greater than 9, were
2778        not recognized as back references, even though there were sufficient
2779        previous subpatterns.
2780    
2781     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2782        versions of gcc, e.g. 2.95.4.
2783    
2784    
2785    Version 6.1 21-Jun-05
2786    ---------------------
2787    
2788     1. There was one reference to the variable "posix" in pcretest.c that was not
2789        surrounded by "#if !defined NOPOSIX".
2790    
2791     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2792        the cross-check on the old pcre_info() function, for the benefit of the
2793        cut-down version of PCRE that is currently imported into Exim.
2794    
2795     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2796        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2797        patterns that start (?i) but I don't think that matters. The use of (?i) is
2798        just an example; this all applies to the other options as well.
2799    
2800     4. Since libtool seems to echo the compile commands it is issuing, the output
2801        from "make" can be reduced a bit by putting "@" in front of each libtool
2802        compile command.
2803    
2804     5. Patch from the folks at Google for configure.in to be a bit more thorough
2805        in checking for a suitable C++ installation before trying to compile the
2806        C++ stuff. This should fix a reported problem when a compiler was present,
2807        but no suitable headers.
2808    
2809     6. The man pages all had just "PCRE" as their title. I have changed them to
2810        be the relevant file name. I have also arranged that these names are
2811        retained in the file doc/pcre.txt, which is a concatenation in text format
2812        of all the man pages except the little individual ones for each function.
2813    
2814     7. The NON-UNIX-USE file had not been updated for the different set of source
2815        files that come with release 6. I also added a few comments about the C++
2816        wrapper.
2817    
2818    
2819    Version 6.0 07-Jun-05
2820    ---------------------
2821    
2822     1. Some minor internal re-organization to help with my DFA experiments.
2823    
2824     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2825        didn't matter for the library itself when fully configured, but did matter
2826        when compiling without UCP support, or within Exim, where the ucp files are
2827        not imported.
2828    
2829     3. Refactoring of the library code to split up the various functions into
2830        different source modules. The addition of the new DFA matching code (see
2831        below) to a single monolithic source would have made it really too
2832        unwieldy, quite apart from causing all the code to be include in a
2833        statically linked application, when only some functions are used. This is
2834        relevant even without the DFA addition now that patterns can be compiled in
2835        one application and matched in another.
2836    
2837        The downside of splitting up is that there have to be some external
2838        functions and data tables that are used internally in different modules of
2839        the library but which are not part of the API. These have all had their
2840        names changed to start with "_pcre_" so that they are unlikely to clash
2841        with other external names.
2842    
2843     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2844        a different (DFA) algorithm. Although it is slower than the original
2845        function, it does have some advantages for certain types of matching
2846        problem.
2847    
2848     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2849        including restarting after a partial match.
2850    
2851     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2852        defined when compiling for Windows was sent to me. I have put it into the
2853        code, though I have no means of testing or verifying it.
2854    
2855     7. Added the pcre_refcount() auxiliary function.
2856    
2857     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2858        match before or at the first newline in the subject string. In pcretest,
2859        the /f option on a pattern can be used to set this.
2860    
2861     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2862        would behave wrongly. This has been present in PCRE since release 4.0.
2863    
2864    10. A number of changes to the pcregrep command:
2865    
2866        (a) Refactored how -x works; insert ^(...)$ instead of setting
2867            PCRE_ANCHORED and checking the length, in preparation for adding
2868            something similar for -w.
2869    
2870        (b) Added the -w (match as a word) option.
2871    
2872        (c) Refactored the way lines are read and buffered so as to have more
2873            than one at a time available.
2874    
2875        (d) Implemented a pcregrep test script.
2876    
2877        (e) Added the -M (multiline match) option. This allows patterns to match
2878            over several lines of the subject. The buffering ensures that at least
2879            8K, or the rest of the document (whichever is the shorter) is available
2880            for matching (and similarly the previous 8K for lookbehind assertions).
2881    
2882        (f) Changed the --help output so that it now says
2883    
2884              -w, --word-regex(p)
2885    
2886            instead of two lines, one with "regex" and the other with "regexp"
2887            because that confused at least one person since the short forms are the
2888            same. (This required a bit of code, as the output is generated
2889            automatically from a table. It wasn't just a text change.)
2890    
2891        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2892            option but starts with a hyphen. Could be a pattern or a path name
2893            starting with a hyphen, for instance.
2894    
2895        (h) "-" can be given as a file name to represent stdin.
2896    
2897        (i) When file names are being printed, "(standard input)" is used for
2898            the standard input, for compatibility with GNU grep. Previously
2899            "<stdin>" was used.
2900    
2901        (j) The option --label=xxx can be used to supply a name to be used for
2902            stdin when file names are being printed. There is no short form.
2903    
2904        (k) Re-factored the options decoding logic because we are going to add
2905            two more options that take data. Such options can now be given in four
2906            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2907    
2908        (l) Added the -A, -B, and -C options for requesting that lines of context
2909            around matches be printed.
2910    
2911        (m) Added the -L option to print the names of files that do not contain
2912            any matching lines, that is, the complement of -l.
2913    
2914        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2915            continue to scan other files.
2916    
2917        (o) The -s option was incorrectly implemented. For compatibility with other
2918            greps, it now suppresses the error message for a non-existent or non-
2919            accessible file (but not the return code). There is a new option called
2920            -q that suppresses the output of matching lines, which was what -s was
2921            previously doing.
2922    
2923        (p) Added --include and --exclude options to specify files for inclusion
2924            and exclusion when recursing.
2925    
2926    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2927        Hopefully, it now does.
2928    
2929    12. Missing cast in pcre_study().
2930    
2931    13. Added an "uninstall" target to the makefile.
2932    
2933    14. Replaced "extern" in the function prototypes in Makefile.in with
2934        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2935        world, but is set differently for Windows.
2936    
2937    15. Added a second compiling function called pcre_compile2(). The only
2938        difference is that it has an extra argument, which is a pointer to an
2939        integer error code. When there is a compile-time failure, this is set
2940        non-zero, in addition to the error test pointer being set to point to an
2941        error message. The new argument may be NULL if no error number is required
2942        (but then you may as well call pcre_compile(), which is now just a
2943        wrapper). This facility is provided because some applications need a
2944        numeric error indication, but it has also enabled me to tidy up the way
2945        compile-time errors are handled in the POSIX wrapper.
2946    
2947    16. Added VPATH=.libs to the makefile; this should help when building with one
2948        prefix path and installing with another. (Or so I'm told by someone who
2949        knows more about this stuff than I do.)
2950    
2951    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2952        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2953        match everything, including newlines. This is not POSIX-compatible, but
2954        somebody wanted the feature. From pcretest it can be activated by using
2955        both the P and the s flags.
2956    
2957    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2958    
2959    19. libpcre.pc was being incorrectly installed as executable.
2960    
2961    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2962        it now also looks for '\r' so that it will work unmodified on Windows.
2963    
2964    21. Added Google's contributed C++ wrapper to the distribution.
2965    
2966    22. Added some untidy missing memory free() calls in pcretest, to keep
2967        Electric Fence happy when testing.
2968    
2969    
2970    
2971    Version 5.0 13-Sep-04
2972    ---------------------
2973    
2974     1. Internal change: literal characters are no longer packed up into items
2975        containing multiple characters in a single byte-string. Each character
2976        is now matched using a separate opcode. However, there may be more than one
2977        byte in the character in UTF-8 mode.
2978    
2979     2. The pcre_callout_block structure has two new fields: pattern_position and
2980        next_item_length. These contain the offset in the pattern to the next match
2981        item, and its length, respectively.
2982    
2983     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2984        insertion of callouts before each pattern item. Added the /C option to
2985        pcretest to make use of this.
2986    
2987     4. On the advice of a Windows user, the lines
2988    
2989          #if defined(_WIN32) || defined(WIN32)
2990          _setmode( _fileno( stdout ), 0x8000 );
2991          #endif  /* defined(_WIN32) || defined(WIN32) */
2992    
2993        have been added to the source of pcretest. This apparently does useful
2994        magic in relation to line terminators.
2995    
2996     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2997        for the benefit of those environments where the "b" makes a difference.
2998    
2999     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3000        to know about it. I have put a hack into configure.in that adds in code
3001        to set GCC=yes if CC=icc. This seems to end up at a point in the
3002        generated configure script that is early enough to affect the setting of
3003        compiler options, which is what is needed, but I have no means of testing
3004        whether it really works. (The user who reported this had patched the
3005        generated configure script, which of course I cannot do.)
3006    
3007        LATER: After change 22 below (new libtool files), the configure script
3008        seems to know about icc (and also ecc). Therefore, I have commented out
3009        this hack in configure.in.
3010    
3011     7. Added support for pkg-config (2 patches were sent in).
3012    
3013     8. Negated POSIX character classes that used a combination of internal tables
3014        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3015        [[:^ascii]]. Typically, they would match almost any characters. The other
3016        POSIX classes were not broken in this way.
3017    
3018     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3019        to find the match, as PCRE was deluded into thinking that the match had to
3020        start at the start point or following a newline. The same bug applied to
3021        patterns with negative forward assertions or any backward assertions
3022        preceding ".*" at the start, unless the pattern required a fixed first
3023        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3024    
3025    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3026        starting at the last subject character, bytes beyond the end of the subject
3027        string were read.
3028    
3029    11. Renamed the variable "class" as "classbits" to make life easier for C++
3030        users. (Previously there was a macro definition, but it apparently wasn't
3031        enough.)
3032    
3033    12. Added the new field "tables" to the extra data so that tables can be passed
3034        in at exec time, or the internal tables can be re-selected. This allows
3035        a compiled regex to be saved and re-used at a later time by a different
3036        program that might have everything at different addresses.
3037    
3038    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3039        -R library as well as a -L library.
3040    
3041    14. The debugging options of pcretest (-d on the command line or D on a
3042        pattern) showed incorrect output for anything following an extended class
3043        that contained multibyte characters and which was followed by a quantifier.
3044    
3045    15. Added optional support for general category Unicode character properties
3046        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3047        support. It adds about 90K to the size of the library. The meanings of the
3048        inbuilt class escapes such as \d and \s have NOT been changed.
3049    
3050    16. Updated pcredemo.c to include calls to free() to release the memory for the
3051        compiled pattern.
3052    
3053    17. The generated file chartables.c was being created in the source directory
3054        instead of in the building directory. This caused the build to fail if the
3055        source directory was different from the building directory, and was
3056        read-only.
3057    
3058    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3059        file. No doubt somebody will tell me if they don't make sense... Also added
3060        Dan Mooney's comments about building on OpenVMS.
3061    
3062    19. Added support for partial matching via the PCRE_PARTIAL option for
3063        pcre_exec() and the \P data escape in pcretest.
3064    
3065    20. Extended pcretest with 3 new pattern features:
3066    
3067        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3068              write the compiled pattern to the file whose name is "rest-of-line".
3069              This is a straight binary dump of the data, with the saved pointer to
3070              the character tables forced to be NULL. The study data, if any, is
3071              written too. After writing, pcretest reads a new pattern.
3072    
3073        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3074              compiled pattern from the given file. There must not be any
3075              occurrences of "<" in the file name (pretty unlikely); if there are,
3076              pcretest will instead treat the initial "<" as a pattern delimiter.
3077              After reading in the pattern, pcretest goes on to read data lines as
3078              usual.
3079    
3080        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3081              and 16-bit fields in a compiled pattern, to simulate a pattern that
3082              was compiled on a host of opposite endianness.
3083    
3084    21. The pcre-exec() function can now cope with patterns that were compiled on
3085        hosts of opposite endianness, with this restriction:
3086    
3087          As for any compiled expression that is saved and used later, the tables
3088          pointer field cannot be preserved; the extra_data field in the arguments
3089          to pcre_exec() should be used to pass in a tables address if a value
3090          other than the default internal tables were used at compile time.
3091    
3092    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3093        now diagnosed as an error. Previously, most of the time, a negative number
3094        would have been treated as zero, but if in addition "ovector" was passed as
3095        NULL, a crash could occur.
3096    
3097    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3098        new versions from the libtool 1.5 distribution (the last one is a copy of
3099        a file called libtool.m4). This seems to have fixed the need to patch
3100        "configure" to support Darwin 1.3 (which I used to do). However, I still
3101        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3102        workstation).
3103    
3104    24. Changed the PCRE licence to be the more standard "BSD" licence.
3105    
3106    
3107    Version 4.5 01-Dec-03
3108    ---------------------
3109    
3110     1. There has been some re-arrangement of the code for the match() function so
3111        that it can be compiled in a version that does not call itself recursively.
3112        Instead, it keeps those local variables that need separate instances for
3113        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3114        needs to "recurse". Keeping track of where control must go is done by means
3115        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3116        hide most of the details from the main code, and operates only if
3117        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3118        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3119        operating.
3120    
3121        To make it easier for callers to provide specially tailored get/free
3122        functions for this usage, two new functions, pcre_stack_malloc, and
3123        pcre_stack_free, are used. They are always called in strict stacking order,
3124        and the size of block requested is always the same.
3125    
3126        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3127        PCRE has been compiled to use the stack or the heap for recursion. The
3128        -C option of pcretest uses this to show which version is compiled.
3129    
3130        A new data escape \S, is added to pcretest; it causes the amounts of store
3131        obtained and freed by both kinds of malloc/free at match time to be added
3132        to the output.
3133    
3134     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3135        what's available on my current Linux desktop machine.
3136    
3137     3. When matching a UTF-8 string, the test for a valid string at the start has
3138        been extended. If start_offset is not zero, PCRE now checks that it points
3139        to a byte that is the start of a UTF-8 character. If not, it returns
3140        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3141        this is necessary because there may be backward assertions in the pattern.
3142        When matching the same subject several times, it may save resources to use
3143        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3144    
3145     4. The code for checking the validity of UTF-8 strings has been tightened so
3146        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3147        containing "overlong sequences".
3148    
3149     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3150        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3151        should have been "&", but it just so happened that all the cases this let
3152        through by mistake were picked up later in the function.
3153    
3154     6. I had used a variable called "isblank" - this is a C99 function, causing
3155        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3156    
3157     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3158        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3159        so the version is shown; (c) stop "make test" echoing "./RunTest".
3160    
3161     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3162    
3163     9. The prototype for memmove() for systems that don't have it was using
3164        size_t, but the inclusion of the header that defines size_t was later. I've
3165        moved the #includes for the C headers earlier to avoid this.
3166    
3167    10. Added some adjustments to the code to make it easier to compiler on certain
3168        special systems:
3169    
3170          (a) Some "const" qualifiers were missing.
3171          (b) Added the macro EXPORT before all exported functions; by default this
3172              is defined to be empty.
3173          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3174              that it reads its output file name as an argument instead of writing
3175              to the standard output and assuming this can be redirected.
3176    
3177    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3178        class containing characters with values greater than 255, PCRE compilation
3179        went into a loop.
3180    
3181    12. A recursive reference to a subpattern that was within another subpattern
3182        that had a minimum quantifier of zero caused PCRE to crash. For example,
3183        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3184        recursion. If the recursively-called subpattern itself had a zero repeat,
3185        that was OK.
3186    
3187    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3188        buffer into which it was copied (for escape processing) was still set at
3189        1024, so long lines caused crashes.
3190    
3191    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3192        "internal error: code overflow...". This applied to any character class
3193        that was followed by a possessive quantifier.
3194    
3195    15. Modified the Makefile to add libpcre.la as a prerequisite for
3196        libpcreposix.la because I was told this is needed for a parallel build to
3197        work.
3198    
3199    16. If a pattern that contained .* following optional items at the start was
3200        studied, the wrong optimizing data was generated, leading to matching
3201        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3202        matching string must start with a or b or c. The correct conclusion for
3203        this pattern is that a match can start with any character.
3204    
3205    
3206    Version 4.4 13-Aug-03
3207    ---------------------
3208    
3209     1. In UTF-8 mode, a character class containing characters with values between
3210        127 and 255 was not handled correctly if the compiled pattern was studied.
3211        In fixing this, I have also improved the studying algorithm for such
3212        classes (slightly).
3213    
3214     2. Three internal functions had redundant arguments passed to them. Removal
3215        might give a very teeny performance improvement.
3216    
3217     3. Documentation bug: the value of the capture_top field in a callout is *one
3218        more than* the number of the hightest numbered captured substring.
3219    
3220     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3221        in incorrectly linking with a previously installed version. They now link
3222        explicitly with libpcre.la.
3223    
3224     5. configure.in no longer needs to recognize Cygwin specially.
3225    
3226     6. A problem in pcre.in for Windows platforms is fixed.
3227    
3228     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3229        pcretest, it used to include the size of the study block as part of its
3230        output. Unfortunately, the structure contains a field that has a different
3231        size on different hardware architectures. This meant that the tests that
3232        showed this size failed. As the block is currently always of a fixed size,
3233        this information isn't actually particularly useful in pcretest output, so
3234        I have just removed it.
3235    
3236     8. Three pre-processor statements accidentally did not start in column 1.
3237        Sadly, there are *still* compilers around that complain, even though
3238        standard C has not required this for well over a decade. Sigh.
3239    
3240     9. In pcretest, the code for checking callouts passed small integers in the
3241        callout_data field, which is a void * field. However, some picky compilers
3242        complained about the casts involved for this on 64-bit systems. Now
3243        pcretest passes the address of the small integer instead, which should get
3244        rid of the warnings.
3245    
3246    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3247        both compile and run time, and gives an error if an invalid UTF-8 sequence
3248        is found. There is a option for disabling this check in cases where the
3249        string is known to be correct and/or the maximum performance is wanted.
3250    
3251    11. In response to a bug report, I changed one line in Makefile.in from
3252    
3253            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3254        to
3255            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3256    
3257        to look similar to other lines, but I have no way of telling whether this
3258        is the right thing to do, as I do not use Windows. No doubt I'll get told
3259        if it's wrong...
3260    
3261    
3262    Version 4.3 21-May-03
3263    ---------------------
3264    
3265    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3266       Makefile.
3267    
3268    2. Some refactoring to improve the quality of the code:
3269    
3270       (i)   The utf8_table... variables are now declared "const".
3271    
3272       (ii)  The code for \cx, which used the "case flipping" table to upper case
3273             lower case letters, now just substracts 32. This is ASCII-specific,
3274             but the whole concept of \cx is ASCII-specific, so it seems
3275             reasonable.
3276    
3277       (iii) PCRE was using its character types table to recognize decimal and
3278             hexadecimal digits in the pattern. This is silly, because it handles
3279             only 0-9, a-f, and A-F, but the character types table is locale-
3280             specific, which means strange things might happen. A private
3281             table is now used for this - though it costs 256 bytes, a table is
3282             much faster than multiple explicit tests. Of course, the standard
3283             character types table is still used for matching digits in subject
3284             strings against \d.
3285    
3286       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3287             ending in _t are). So I've renamed it as ESC_tee.
3288    
3289    3. The first argument for regexec() in the POSIX wrapper should have been
3290       defined as "const".
3291    
3292    4. Changed pcretest to use malloc() for its buffers so that they can be
3293       Electric Fenced for debugging.
3294    
3295    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3296       to read one or more bytes before the start of the subject string. Often this
3297       had no effect on PCRE's behaviour, but in some circumstances it could
3298       provoke a segmentation fault.
3299    
3300    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3301       to try to read one or more bytes before the start of the subject string.
3302    
3303    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3304       UTF-8 support could misbehave in various ways if the subject string
3305       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3306       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3307       back over UTF-8 characters.)
3308    
3309    
3310    Version 4.2 14-Apr-03
3311    ---------------------
3312    
3313    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3314    
3315    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3316         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3317         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3318         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3319         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3320           and BUILD_EXEEXT
3321         Note: automatic setting of the BUILD variables is not yet working
3322         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3323           compile-time but not at link-time
3324         [LINK]: use for linking executables only
3325         make different versions for Windows and non-Windows
3326         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3327           libraries
3328         [LINK_FOR_BUILD]: new variable
3329         [OBJEXT]: use throughout
3330         [EXEEXT]: use throughout
3331         <winshared>: new target
3332         <wininstall>: new target
3333         <dftables.o>: use native compiler
3334         <dftables>: use native linker
3335         <install>: handle Windows platform correctly
3336         <clean>: ditto
3337         <check>: ditto
3338         copy DLL to top builddir before testing
3339    
3340       As part of these changes, -no-undefined was removed again. This was reported
3341       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3342       in any case.
3343    
3344    3. Some tidies to get rid of compiler warnings:
3345    
3346       . In the match_data structure, match_limit was an unsigned long int, whereas
3347         match_call_count was an int. I've made them both unsigned long ints.
3348    
3349       . In pcretest the fact that a const uschar * doesn't automatically cast to
3350         a void * provoked a warning.
3351    
3352       . Turning on some more compiler warnings threw up some "shadow" variables
3353         and a few more missing casts.
3354    
3355    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3356       option, a class that contained a single character with a value between 128
3357       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3358    
3359    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3360       option, a class that contained several characters, but with at least one
3361       whose value was between 128 and 255 caused PCRE to crash.
3362    
3363    
3364    Version 4.1 12-Mar-03
3365    ---------------------
3366    
3367    1. Compiling with gcc -pedantic found a couple of places where casts were
3368    needed, and a string in dftables.c that was longer than standard compilers are
3369    required to support.
3370    
3371    2. Compiling with Sun's compiler found a few more places where the code could
3372    be tidied up in order to avoid warnings.
3373    
3374    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3375    first of these names is deprecated in the latest Autoconf in favour of the name
3376    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3377    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3378    analogy I have changed it to CFLAGS_FOR_BUILD.
3379    
3380    4. Added -no-undefined to the linking command in the Makefile, because this is
3381    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3382    linking step for the pcreposix library.
3383    
3384    5. PCRE was failing to diagnose the case of two named groups with the same
3385    name.
3386    
3387    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3388    literal character that is needed in the subject for a match, and scans along to
3389    ensure that it is present before embarking on the full matching process. This
3390    saves time in cases of nested unlimited repeats that are never going to match.
3391    Problem: the scan can take a lot of time if the subject is very long (e.g.
3392    megabytes), thus penalizing straightforward matches. It is now done only if the
3393    amount of subject to be scanned is less than 1000 bytes.
3394    
3395    7. A lesser problem with the same optimization is that it was recording the
3396    first character of an anchored pattern as "needed", thus provoking a search
3397    right along the subject, even when the first match of the pattern was going to
3398    fail. The "needed" character is now not set for anchored patterns, unless it
3399    follows something in the pattern that is of non-fixed length. Thus, it still
3400    fulfils its original purpose of finding quick non-matches in cases of nested
3401    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3402    
3403    
3404    Version 4.0 17-Feb-03
3405    ---------------------
3406    
3407    1. If a comment in an extended regex that started immediately after a meta-item
3408    extended to the end of string, PCRE compiled incorrect data. This could lead to
3409    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3410    
3411    2. Moved to autoconf 2.53 and libtool 1.4.2.
3412    
3413    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3414    the special perltest8 script is no longer needed - all the tests can be run
3415    from a single perltest script.
3416    
3417    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3418    by \s. It has now been removed in PCRE. This means it isn't recognized as
3419    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3420    class [:space:] *does* include VT, thereby creating a mess.
3421    
3422    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3423    space and tab.
3424    
3425    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3426    its new features into the main test script, reducing the number of scripts.
3427    
3428    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3429    were backward compatible, and made the (?i) apply to the whole pattern, as if
3430    /i were given. Now it behaves more logically, and applies the option setting
3431    only to what follows. PCRE has been changed to follow suit. However, if it
3432    finds options settings right at the start of the pattern, it extracts them into
3433    the global options, as before. Thus, they show up in the info data.
3434    
3435    8. Added support for the \Q...\E escape sequence. Characters in between are
3436    treated as literals. This is slightly different from Perl in that $ and @ are
3437    also handled as literals inside the quotes. In Perl, they will cause variable
3438    interpolation. Note the following examples:
3439    
3440        Pattern            PCRE matches      Perl matches
3441    
3442        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3443        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3444        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3445    
3446    For compatibility with Perl, \Q...\E sequences are recognized inside character
3447    classes as well as outside them.
3448    
3449    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3450    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3451    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3452    signed/unsigned warnings.
3453    
3454    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3455    option for pcretest, so I've replaced it by a simple function that does just
3456    that job.
3457    
3458    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3459    "pcregrep -".
3460    
3461    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3462    Java package. This provides some syntactic sugar for simple cases of what my
3463    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3464    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3465    item, you can use this simplified notation. Note that only makes sense with
3466    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3467    greediness, whatever the setting of the PCRE_UNGREEDY option.
3468    
3469    13. A change of greediness default within a pattern was not taking effect at
3470    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3471    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3472    was abstracted outside.
3473    
3474    14. PCRE now supports the \G assertion. It is true when the current matching
3475    position is at the start point of the match. This differs from \A when the
3476    starting offset is non-zero. Used with the /g option of pcretest (or similar
3477    code), it works in the same way as it does for Perl's /g option. If all
3478    alternatives of a regex begin with \G, the expression is anchored to the start
3479    match position, and the "anchored" flag is set in the compiled expression.
3480    
3481    15. Some bugs concerning the handling of certain option changes within patterns
3482    have been fixed. These applied to options other than (?ims). For example,
3483    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3484    been the other way round. Some of this was related to change 7 above.
3485    
3486    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3487    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3488    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3489    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3490    
3491    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3492    mode. Unlike ".", it always matches newline, whatever the setting of
3493    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3494    assertions. Perl allows it, but it doesn't (in general) work because it can't
3495    calculate the length of the lookbehind. At least, that's the case for Perl
3496    5.8.0 - I've been told they are going to document that it doesn't work in
3497    future.
3498    
3499    18. Added an error diagnosis for escapes that PCRE does not support: these are
3500    \L, \l, \N, \P, \p, \U, \u, and \X.
3501    
3502    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3503    reading past the end of the pattern in cases such as /[abcd/.
3504    
3505    20. PCRE was getting more memory than necessary for patterns with classes that
3506    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3507    
3508    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3509    compiling PCRE for use with Virtual Pascal.
3510    
3511    22. Small fix to the Makefile to make it work properly if the build is done
3512    outside the source tree.
3513    
3514    23. Added a new extension: a condition to go with recursion. If a conditional
3515    subpattern starts with (?(R) the "true" branch is used if recursion has
3516    happened, whereas the "false" branch is used only at the top level.
3517    
3518    24. When there was a very long string of literal characters (over 255 bytes
3519    without UTF support, over 250 bytes with UTF support), the computation of how
3520    much memory was required could be incorrect, leading to segfaults or other
3521    strange effects.
3522    
3523    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3524    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3525    there was a subsequent back reference to those brackets. This meant that, for
3526    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3527    possible to check for precisely this case. All we can do is abandon the
3528    optimization if .* occurs inside capturing brackets when there are any back
3529    references whatsoever. (See below for a better fix that came later.)
3530    
3531    26. The handling of the optimization for finding the first character of a
3532    non-anchored pattern, and for finding a character that is required later in the
3533    match were failing in some cases. This didn't break the matching; it just
3534    failed to optimize when it could. The way this is done has been re-implemented.
3535    
3536    27. Fixed typo in error message for invalid (?R item (it said "(?p").