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

Diff of /code/trunk/ChangeLog

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

revision 99 by ph10, Tue Mar 6 12:27:42 2007 UTC revision 753 by ph10, Sat Nov 19 17:03:35 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 05-Mar-07  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    
54    Version 8.20 21-Oct-2011
55    ------------------------
56    
57    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
58        a POSIX class. After further experiments with Perl, which convinced me that
59        Perl has bugs and confusions, a closing square bracket is no longer allowed
60        in a POSIX name. This bug also affected patterns with classes that started
61        with full stops.
62    
63    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
64        captured substring, but while checking the failing first alternative,
65        substring 1 is temporarily captured. If the output vector supplied to
66        pcre_exec() was not big enough for this capture, the yield of the function
67        was still zero ("insufficient space for captured substrings"). This cannot
68        be totally fixed without adding another stack variable, which seems a lot
69        of expense for a edge case. However, I have improved the situation in cases
70        such as /(a)(b)x|abc/ matched against "abc", where the return code
71        indicates that fewer than the maximum number of slots in the ovector have
72        been set.
73    
74    3.  Related to (2) above: when there are more back references in a pattern than
75        slots in the output vector, pcre_exec() uses temporary memory during
76        matching, and copies in the captures as far as possible afterwards. It was
77        using the entire output vector, but this conflicts with the specification
78        that only 2/3 is used for passing back captured substrings. Now it uses
79        only the first 2/3, for compatibility. This is, of course, another edge
80        case.
81    
82    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
83        main code base, and can be used by building with --enable-jit. When this is
84        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
85        runtime --no-jit option is given.
86    
87    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
88        ovector, the return from the function was zero, implying that there were
89        other matches that did not fit. The correct "exactly full" value is now
90        returned.
91    
92    6.  If a subpattern that was called recursively or as a subroutine contained
93        (*PRUNE) or any other control that caused it to give a non-standard return,
94        invalid errors such as "Error -26 (nested recursion at the same subject
95        position)" or even infinite loops could occur.
96    
97    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
98        computing the minimum length on reaching *ACCEPT, and so ended up with the
99        wrong value of 1 rather than 0. Further investigation indicates that
100        computing a minimum subject length in the presence of *ACCEPT is difficult
101        (think back references, subroutine calls), and so I have changed the code
102        so that no minimum is registered for a pattern that contains *ACCEPT.
103    
104    8.  If (*THEN) was present in the first (true) branch of a conditional group,
105        it was not handled as intended. [But see 16 below.]
106    
107    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
108        Sheri Pierce.
109    
110    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
111        the first byte in a match must be "a".
112    
113    11. Change 17 for 8.13 increased the recursion depth for patterns like
114        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
115        pattern contains any instances of (*THEN). If it does not, the old
116        optimizations are restored. It would be nice to do this on a per-group
117        basis, but at the moment that is not feasible.
118    
119    12. In some environments, the output of pcretest -C is CRLF terminated. This
120        broke RunTest's code that checks for the link size. A single white space
121        character after the value is now allowed for.
122    
123    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
124        For "fr", it uses the Windows-specific input and output files.
125    
126    14. If (*THEN) appeared in a group that was called recursively or as a
127        subroutine, it did not work as intended. [But see next item.]
128    
129    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
130        pattern fragments (but not containing any | characters). If A and B are
131        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
132        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
133        D. In other words, Perl considers parentheses that do not contain any |
134        characters to be part of a surrounding alternative, whereas PCRE was
135        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
136        differently. PCRE now behaves in the same way as Perl, except in the case
137        of subroutine/recursion calls such as (?1) which have in any case always
138        been different (but PCRE had them first :-).
139    
140    16. Related to 15 above: Perl does not treat the | in a conditional group as
141        creating alternatives. Such a group is treated in the same way as an
142        ordinary group without any | characters when processing (*THEN). PCRE has
143        been changed to match Perl's behaviour.
144    
145    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
146        RunGrepTest script failed.
147    
148    18. Change 22 for version 13 caused atomic groups to use more stack. This is
149        inevitable for groups that contain captures, but it can lead to a lot of
150        stack use in large patterns. The old behaviour has been restored for atomic
151        groups that do not contain any capturing parentheses.
152    
153    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
154        suppress the check for a minimum subject length at run time. (If it was
155        given to pcre_exec() or pcre_dfa_exec() it did work.)
156    
157    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
158        fail to work when decoding hex characters in data strings in EBCDIC
159        environments.
160    
161    21. It appears that in at least one Mac OS environment, the isxdigit() function
162        is implemented as a macro that evaluates to its argument more than once,
163        contravening the C 90 Standard (I haven't checked a later standard). There
164        was an instance in pcretest which caused it to go wrong when processing
165        \x{...} escapes in subject strings. The has been rewritten to avoid using
166        things like p++ in the argument of isxdigit().
167    
168    
169    Version 8.13 16-Aug-2011
170    ------------------------
171    
172    1.  The Unicode data tables have been updated to Unicode 6.0.0.
173    
174    2.  Two minor typos in pcre_internal.h have been fixed.
175    
176    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
177        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
178        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
179    
180    4.  There were a number of related bugs in the code for matching backrefences
181        caselessly in UTF-8 mode when codes for the characters concerned were
182        different numbers of bytes. For example, U+023A and U+2C65 are an upper
183        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
184        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
185        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
186        2-byte code at the end of the subject (it thought there wasn't enough data
187        left).
188    
189    5.  Comprehensive information about what went wrong is now returned by
190        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
191        as the output vector has at least 2 elements. The offset of the start of
192        the failing character and a reason code are placed in the vector.
193    
194    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
195        now returned is for the first byte of the failing character, instead of the
196        last byte inspected. This is an incompatible change, but I hope it is small
197        enough not to be a problem. It makes the returned offset consistent with
198        pcre_exec() and pcre_dfa_exec().
199    
200    7.  pcretest now gives a text phrase as well as the error number when
201        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
202        failure, the offset and reason code are output.
203    
204    8.  When \R was used with a maximizing quantifier it failed to skip backwards
205        over a \r\n pair if the subsequent match failed. Instead, it just skipped
206        back over a single character (\n). This seems wrong (because it treated the
207        two characters as a single entity when going forwards), conflicts with the
208        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
209        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
210        has been changed.
211    
212    9.  Some internal refactoring has changed the processing so that the handling
213        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
214        time (the PCRE_DOTALL option was changed this way some time ago: version
215        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
216        which was always a bit of a fudge. It also means that there is one less
217        argument for the match() function, which reduces its stack requirements
218        slightly. This change also fixes an incompatibility with Perl: the pattern
219        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
220    
221    10. More internal refactoring has drastically reduced the number of recursive
222        calls to match() for possessively repeated groups such as (abc)++ when
223        using pcre_exec().
224    
225    11. While implementing 10, a number of bugs in the handling of groups were
226        discovered and fixed:
227    
228        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
229        (a|)*(?1) gave a compile-time internal error.
230        ((a|)+)+  did not notice that the outer group could match an empty string.
231        (^a|^)+   was not marked as anchored.
232        (.*a|.*)+ was not marked as matching at start or after a newline.
233    
234    12. Yet more internal refactoring has removed another argument from the match()
235        function. Special calls to this function are now indicated by setting a
236        value in a variable in the "match data" data block.
237    
238    13. Be more explicit in pcre_study() instead of relying on "default" for
239        opcodes that mean there is no starting character; this means that when new
240        ones are added and accidentally left out of pcre_study(), testing should
241        pick them up.
242    
243    14. The -s option of pcretest has been documented for ages as being an old
244        synonym of -m (show memory usage). I have changed it to mean "force study
245        for every regex", that is, assume /S for every regex. This is similar to -i
246        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
247        using it. It makes it easier to run collections of tests with and without
248        study enabled, and thereby test pcre_study() more easily. All the standard
249        tests are now run with and without -s (but some patterns can be marked as
250        "never study" - see 20 below).
251    
252    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
253        restoration of the capturing data to the outer values was not happening
254        correctly.
255    
256    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
257        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
258        pattern had matched an empty string, and so incorrectly returned a no
259        match.
260    
261    17. There was optimizing code for the last branch of non-capturing parentheses,
262        and also for the obeyed branch of a conditional subexpression, which used
263        tail recursion to cut down on stack usage. Unfortunately, now that there is
264        the possibility of (*THEN) occurring in these branches, tail recursion is
265        no longer possible because the return has to be checked for (*THEN). These
266        two optimizations have therefore been removed. [But see 8.20/11 above.]
267    
268    18. If a pattern containing \R was studied, it was assumed that \R always
269        matched two bytes, thus causing the minimum subject length to be
270        incorrectly computed because \R can also match just one byte.
271    
272    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
273        was incorrectly computed.
274    
275    20. If /S is present twice on a test pattern in pcretest input, it now
276        *disables* studying, thereby overriding the use of -s on the command line
277        (see 14 above). This is necessary for one or two tests to keep the output
278        identical in both cases.
279    
280    21. When (*ACCEPT) was used in an assertion that matched an empty string and
281        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
282    
283    22. When an atomic group that contained a capturing parenthesis was
284        successfully matched, but the branch in which it appeared failed, the
285        capturing was not being forgotten if a higher numbered group was later
286        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
287        group 1 to "a", when in fact it should be unset. This applied to multi-
288        branched capturing and non-capturing groups, repeated or not, and also to
289        positive assertions (capturing in negative assertions does not happen
290        in PCRE) and also to nested atomic groups.
291    
292    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
293        subject after a captured substring, to make it easier to tell which of a
294        number of identical substrings has been captured.
295    
296    24. The way atomic groups are processed by pcre_exec() has been changed so that
297        if they are repeated, backtracking one repetition now resets captured
298        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
299        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
300        "aaa". Previously, it would have been "a". As part of this code
301        refactoring, the way recursive calls are handled has also been changed.
302    
303    25. If an assertion condition captured any substrings, they were not passed
304        back unless some other capturing happened later. For example, if
305        (?(?=(a))a) was matched against "a", no capturing was returned.
306    
307    26. When studying a pattern that contained subroutine calls or assertions,
308        the code for finding the minimum length of a possible match was handling
309        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
310        group 1 called group 2 while simultaneously a separate group 2 called group
311        1). A stack overflow occurred in this case. I have fixed this by limiting
312        the recursion depth to 10.
313    
314    27. Updated RunTest.bat in the distribution to the version supplied by Tom
315        Fortmann. This supports explicit test numbers on the command line, and has
316        argument validation and error reporting.
317    
318    28. An instance of \X with an unlimited repeat could fail if at any point the
319        first character it looked at was a mark character.
320    
321    29. Some minor code refactoring concerning Unicode properties and scripts
322        should reduce the stack requirement of match() slightly.
323    
324    30. Added the '=' option to pcretest to check the setting of unused capturing
325        slots at the end of the pattern, which are documented as being -1, but are
326        not included in the return count.
327    
328    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
329        compiled something random. Now it gives a compile-time error (as does
330        Perl).
331    
332    32. A *MARK encountered during the processing of a positive assertion is now
333        recorded and passed back (compatible with Perl).
334    
335    33. If --only-matching or --colour was set on a pcregrep call whose pattern
336        had alternative anchored branches, the search for a second match in a line
337        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
338        matched the line "0102" twice. The same bug affected patterns that started
339        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
340        twice.
341    
342    34. Previously, PCRE did not allow quantification of assertions. However, Perl
343        does, and because of capturing effects, quantifying parenthesized
344        assertions may at times be useful. Quantifiers are now allowed for
345        parenthesized assertions.
346    
347    35. A minor code tidy in pcre_compile() when checking options for \R usage.
348    
349    36. \g was being checked for fancy things in a character class, when it should
350        just be a literal "g".
351    
352    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
353        appearance of a nested POSIX class supersedes an apparent external class.
354        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
355        unescaped square brackets may also appear as part of class names. For
356        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
357        more like Perl. (But see 8.20/1 above.)
358    
359    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
360        was because it thought it was \N{name}, which is not supported).
361    
362    39. Add minix to OS list not supporting the -S option in pcretest.
363    
364    40. PCRE tries to detect cases of infinite recursion at compile time, but it
365        cannot analyze patterns in sufficient detail to catch mutual recursions
366        such as ((?1))((?2)). There is now a runtime test that gives an error if a
367        subgroup is called recursively as a subpattern for a second time at the
368        same position in the subject string. In previous releases this might have
369        been caught by the recursion limit, or it might have run out of stack.
370    
371    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
372        happen only once. PCRE was, however incorrectly giving a compile time error
373        "recursive call could loop indefinitely" because it cannot analyze the
374        pattern in sufficient detail. The compile time test no longer happens when
375        PCRE is compiling a conditional subpattern, but actual runaway loops are
376        now caught at runtime (see 40 above).
377    
378    42. It seems that Perl allows any characters other than a closing parenthesis
379        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
380        has been changed to be the same.
381    
382    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
383        as not to get warnings when autogen.sh is called. Also changed
384        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
385    
386    44. To help people who use pcregrep to scan files containing exceedingly long
387        lines, the following changes have been made:
388    
389        (a) The default value of the buffer size parameter has been increased from
390            8K to 20K. (The actual buffer used is three times this size.)
391    
392        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
393            PCRE is built.
394    
395        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
396            to be set at run time.
397    
398        (d) Numerical values in pcregrep options can be followed by K or M, for
399            example --buffer-size=50K.
400    
401        (e) If a line being scanned overflows pcregrep's buffer, an error is now
402            given and the return code is set to 2.
403    
404    45. Add a pointer to the latest mark to the callout data block.
405    
406    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
407        partial match of an empty string instead of no match. This was specific to
408        the use of ".".
409    
410    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
411        complete match instead of a partial match. This bug was dependent on both
412        the PCRE_UTF8 and PCRE_DOTALL options being set.
413    
414    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
415        starting byte set, because \b was not being ignored.
416    
417    
418    Version 8.12 15-Jan-2011
419    ------------------------
420    
421    1.  Fixed some typos in the markup of the man pages, and wrote a script that
422        checks for such things as part of the documentation building process.
423    
424    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
425        --match-limit and --recursion-limit options (added for 8.11). In
426        particular, this made one of the standard tests fail. (The integer value
427        went into the wrong half of a long int.)
428    
429    3.  If the --colour option was given to pcregrep with -v (invert match), it
430        did strange things, either producing crazy output, or crashing. It should,
431        of course, ignore a request for colour when reporting lines that do not
432        match.
433    
434    4.  Another pcregrep bug caused similar problems if --colour was specified with
435        -M (multiline) and the pattern match finished with a line ending.
436    
437    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
438        matched in multiline mode, the following line was shown as part of the
439        match. This seems wrong, so I have changed it.
440    
441    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
442        the check for further matches in the same line (so they could be coloured)
443        to overrun the end of the current line. If another match was found, it was
444        incorrectly shown (and then shown again when found in the next line).
445    
446    7.  If pcregrep was compiled under Windows, there was a reference to the
447        function pcregrep_exit() before it was defined. I am assuming this was
448        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
449        reported by a user. I've moved the definition above the reference.
450    
451    
452    Version 8.11 10-Dec-2010
453    ------------------------
454    
455    1.  (*THEN) was not working properly if there were untried alternatives prior
456        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
457        backtracked to try for "b" instead of moving to the next alternative branch
458        at the same level (in this case, to look for "c"). The Perl documentation
459        is clear that when (*THEN) is backtracked onto, it goes to the "next
460        alternative in the innermost enclosing group".
461    
462    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
463        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
464        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
465        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
466        (*THEN).
467    
468    3.  If \s appeared in a character class, it removed the VT character from
469        the class, even if it had been included by some previous item, for example
470        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
471        of \s, but is part of the POSIX "space" class.)
472    
473    4.  A partial match never returns an empty string (because you can always
474        match an empty string at the end of the subject); however the checking for
475        an empty string was starting at the "start of match" point. This has been
476        changed to the "earliest inspected character" point, because the returned
477        data for a partial match starts at this character. This means that, for
478        example, /(?<=abc)def/ gives a partial match for the subject "abc"
479        (previously it gave "no match").
480    
481    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
482        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
483        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
484        has an implication that the given string is incomplete (because a partial
485        match is preferred over a full match). For this reason, these items now
486        give a partial match in this situation. [Aside: previously, the one case
487        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
488        match rather than a full match, which was wrong by the old rules, but is
489        now correct.]
490    
491    6.  There was a bug in the handling of #-introduced comments, recognized when
492        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
493        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
494        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
495        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
496        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
497        places in pcre_compile().
498    
499    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
500        comments when looking ahead for named forward references to subpatterns,
501        the only newline sequence it recognized was NL. It now handles newlines
502        according to the set newline convention.
503    
504    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
505        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
506        cater for a lack of strerror(). These oversights have been fixed.
507    
508    9.  Added --match-limit and --recursion-limit to pcregrep.
509    
510    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
511    
512    11. When the -o option was used, pcregrep was setting a return code of 1, even
513        when matches were found, and --line-buffered was not being honoured.
514    
515    12. Added an optional parentheses number to the -o and --only-matching options
516        of pcregrep.
517    
518    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
519        can match an empty string. The code to do it in pcretest and pcredemo
520        needed fixing:
521    
522        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
523            only one byte after an empty string match just before CRLF (this case
524            just got forgotten; "any" and "anycrlf" were OK).
525    
526        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
527            mode when an empty string match preceded an ASCII character followed by
528            a non-ASCII character. (The code for advancing by one character rather
529            than one byte was nonsense.)
530    
531        (c) The pcredemo.c sample program did not have any code at all to handle
532            the cases when CRLF is a valid newline sequence.
533    
534    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
535        as a starting offset was within the subject string. There is now a new
536        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
537        negative or greater than the length of the string. In order to test this,
538        pcretest is extended to allow the setting of negative starting offsets.
539    
540    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
541        starting offset points to the beginning of a UTF-8 character was
542        unnecessarily clumsy. I tidied it up.
543    
544    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
545        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
546    
547    17. Nobody had reported that the --include_dir option, which was added in
548        release 7.7 should have been called --include-dir (hyphen, not underscore)
549        for compatibility with GNU grep. I have changed it to --include-dir, but
550        left --include_dir as an undocumented synonym, and the same for
551        --exclude-dir, though that is not available in GNU grep, at least as of
552        release 2.5.4.
553    
554    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
555        characters from a string of bytes) have been redefined so as not to use
556        loops, in order to improve performance in some environments. At the same
557        time, I abstracted some of the common code into auxiliary macros to save
558        repetition (this should not affect the compiled code).
559    
560    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
561        compile-time error is now given if \c is not followed by an ASCII
562        character, that is, a byte less than 128. (In EBCDIC mode, the code is
563        different, and any byte value is allowed.)
564    
565    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
566        START_OPTIMIZE option, which is now allowed at compile time - but just
567        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
568        to pcregrep and other applications that have no direct access to PCRE
569        options. The new /Y option in pcretest sets this option when calling
570        pcre_compile().
571    
572    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
573        back references. Groups containing recursive back references were forced to
574        be atomic by that change, but in the case of named groups, the amount of
575        memory required was incorrectly computed, leading to "Failed: internal
576        error: code overflow". This has been fixed.
577    
578    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
579        pcretest.c, to avoid build problems in some Borland environments.
580    
581    
582    Version 8.10 25-Jun-2010
583    ------------------------
584    
585    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
586        THEN.
587    
588    2.  (*ACCEPT) was not working when inside an atomic group.
589    
590    3.  Inside a character class, \B is treated as a literal by default, but
591        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
592        causes the error). The code is unchanged, but I tidied the documentation.
593    
594    4.  Inside a character class, PCRE always treated \R and \X as literals,
595        whereas Perl faults them if its -w option is set. I have changed PCRE so
596        that it faults them when PCRE_EXTRA is set.
597    
598    5.  Added support for \N, which always matches any character other than
599        newline. (It is the same as "." when PCRE_DOTALL is not set.)
600    
601    6.  When compiling pcregrep with newer versions of gcc which may have
602        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
603        declared with attribute warn_unused_result" were given. Just casting the
604        result to (void) does not stop the warnings; a more elaborate fudge is
605        needed. I've used a macro to implement this.
606    
607    7.  Minor change to pcretest.c to avoid a compiler warning.
608    
609    8.  Added four artifical Unicode properties to help with an option to make
610        \s etc use properties (see next item). The new properties are: Xan
611        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
612    
613    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
614        use Unicode properties. (*UCP) at the start of a pattern can be used to set
615        this option. Modified pcretest to add /W to test this facility. Added
616        REG_UCP to make it available via the POSIX interface.
617    
618    10. Added --line-buffered to pcregrep.
619    
620    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
621        studied, and the match started with a letter with a code point greater than
622        127 whose first byte was different to the first byte of the other case of
623        the letter, the other case of this starting letter was not recognized
624        (#976).
625    
626    12. If a pattern that was studied started with a repeated Unicode property
627        test, for example, \p{Nd}+, there was the theoretical possibility of
628        setting up an incorrect bitmap of starting bytes, but fortunately it could
629        not have actually happened in practice until change 8 above was made (it
630        added property types that matched character-matching opcodes).
631    
632    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
633        possible starting bytes for non-anchored patterns.
634    
635    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
636        \R, and also a number of cases that involve Unicode properties, both
637        explicit and implicit when PCRE_UCP is set.
638    
639    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
640        input, it could crash or give wrong results if characters with values
641        greater than 0xc0 were present in the subject string. (Detail: it assumed
642        UTF-8 input when processing these items.)
643    
644    16. Added a lot of (int) casts to avoid compiler warnings in systems where
645        size_t is 64-bit (#991).
646    
647    17. Added a check for running out of memory when PCRE is compiled with
648        --disable-stack-for-recursion (#990).
649    
650    18. If the last data line in a file for pcretest does not have a newline on
651        the end, a newline was missing in the output.
652    
653    19. The default pcre_chartables.c file recognizes only ASCII characters (values
654        less than 128) in its various bitmaps. However, there is a facility for
655        generating tables according to the current locale when PCRE is compiled. It
656        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
657        characters, are recognized by isspace() and therefore were getting set in
658        these tables, and indeed these tables seem to approximate to ISO 8859. This
659        caused a problem in UTF-8 mode when pcre_study() was used to create a list
660        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
661        which of course cannot start UTF-8 characters. I have changed the code so
662        that only real ASCII characters (less than 128) and the correct starting
663        bytes for UTF-8 encodings are set for characters greater than 127 when in
664        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
665        altogether.)
666    
667    20. Added the /T option to pcretest so as to be able to run tests with non-
668        standard character tables, thus making it possible to include the tests
669        used for 19 above in the standard set of tests.
670    
671    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
672        reference to a subpattern the other side of a comment that contains an
673        opening parenthesis caused either an internal compiling error, or a
674        reference to the wrong subpattern.
675    
676    
677    Version 8.02 19-Mar-2010
678    ------------------------
679    
680    1.  The Unicode data tables have been updated to Unicode 5.2.0.
681    
682    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
683        configured.
684    
685    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
686        original author of that file, following a query about its status.
687    
688    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
689        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
690    
691    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
692        quantifier applied to a forward-referencing subroutine call, could compile
693        incorrect code or give the error "internal error: previously-checked
694        referenced subpattern not found".
695    
696    6.  Both MS Visual Studio and Symbian OS have problems with initializing
697        variables to point to external functions. For these systems, therefore,
698        pcre_malloc etc. are now initialized to local functions that call the
699        relevant global functions.
700    
701    7.  There were two entries missing in the vectors called coptable and poptable
702        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
703        I've fixed the data, and added a kludgy way of testing at compile time that
704        the lengths are correct (equal to the number of opcodes).
705    
706    8.  Following on from 7, I added a similar kludge to check the length of the
707        eint vector in pcreposix.c.
708    
709    9.  Error texts for pcre_compile() are held as one long string to avoid too
710        much relocation at load time. To find a text, the string is searched,
711        counting zeros. There was no check for running off the end of the string,
712        which could happen if a new error number was added without updating the
713        string.
714    
715    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
716    
717    11. \K was not working if it appeared in an atomic group or in a group that
718        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
719        \K is "not well defined" if used in an assertion. PCRE now accepts it if
720        the assertion is positive, but not if it is negative.
721    
722    12. Change 11 fortuitously reduced the size of the stack frame used in the
723        "match()" function of pcre_exec.c by one pointer. Forthcoming
724        implementation of support for (*MARK) will need an extra pointer on the
725        stack; I have reserved it now, so that the stack frame size does not
726        decrease.
727    
728    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
729        item in branch that calls a recursion is a subroutine call - as in the
730        second branch in the above example - was incorrectly given the compile-
731        time error "recursive call could loop indefinitely" because pcre_compile()
732        was not correctly checking the subroutine for matching a non-empty string.
733    
734    14. The checks for overrunning compiling workspace could trigger after an
735        overrun had occurred. This is a "should never occur" error, but it can be
736        triggered by pathological patterns such as hundreds of nested parentheses.
737        The checks now trigger 100 bytes before the end of the workspace.
738    
739    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
740    
741    
742    Version 8.01 19-Jan-2010
743    ------------------------
744    
745    1.  If a pattern contained a conditional subpattern with only one branch (in
746        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
747        computed the wrong minimum data length (which is of course zero for such
748        subpatterns). This could cause incorrect "no match" results.
749    
750    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
751        the pattern is reset in the first branch, pcre_compile() failed with
752        "internal error: code overflow at offset...". This happened only when
753        the reset was to the original external option setting. (An optimization
754        abstracts leading options settings into an external setting, which was the
755        cause of this.)
756    
757    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
758        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
759        assertion pattern did not match (meaning that the assertion was true), it
760        was incorrectly treated as false if the SKIP had been reached during the
761        matching. This also applied to assertions used as conditions.
762    
763    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
764        assertion subpattern, including such a pattern used as a condition,
765        unpredictable results occurred, instead of the error return
766        PCRE_ERROR_DFA_UITEM.
767    
768    5.  The C++ GlobalReplace function was not working like Perl for the special
769        situation when an empty string is matched. It now does the fancy magic
770        stuff that is necessary.
771    
772    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
773        removed. (These were left over from very, very early versions of PCRE.)
774    
775    7.  Some cosmetic changes to the code to make life easier when compiling it
776        as part of something else:
777    
778        (a) Change DEBUG to PCRE_DEBUG.
779    
780        (b) In pcre_compile(), rename the member of the "branch_chain" structure
781            called "current" as "current_branch", to prevent a collision with the
782            Linux macro when compiled as a kernel module.
783    
784        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
785            prevent a collision with the Linux macro when compiled as a kernel
786            module.
787    
788    8.  In pcre_compile() there are some checks for integer overflows that used to
789        cast potentially large values to (double). This has been changed to that
790        when building, a check for int64_t is made, and if it is found, it is used
791        instead, thus avoiding the use of floating point arithmetic. (There is no
792        other use of FP in PCRE.) If int64_t is not found, the fallback is to
793        double.
794    
795    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
796        2005 (difference between two addresses compared to an unsigned value).
797    
798    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
799        custom one, because of the following reported problem in Windows:
800    
801          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
802              under Win32.
803          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
804              therefore missing the function definition.
805          - The compiler thus generates a "C" signature for the test function.
806          - The linker fails to find the "C" function.
807          - PCRE fails to configure if asked to do so against libbz2.
808    
809    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
810        messages were output:
811    
812          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
813          rerunning libtoolize, to keep the correct libtool macros in-tree.
814          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
815    
816        I have done both of these things.
817    
818    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
819        most of the time, it *can* run out if it is given a pattern that contains a
820        runaway infinite recursion. I updated the discussion in the pcrestack man
821        page.
822    
823    13. Now that we have gone to the x.xx style of version numbers, the minor
824        version may start with zero. Using 08 or 09 is a bad idea because users
825        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
826        interpreted as invalid octal numbers. I've updated the previous comment in
827        configure.ac, and also added a check that gives an error if 08 or 09 are
828        used.
829    
830    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
831        causing partial matching to fail when the end of the subject matched \W
832        in a UTF-8 pattern where \W was quantified with a minimum of 3.
833    
834    15. There were some discrepancies between the declarations in pcre_internal.h
835        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
836        their definitions. The declarations used "const uschar *" and the
837        definitions used USPTR. Even though USPTR is normally defined as "const
838        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
839        reported that: "This difference in casting confuses some C++ compilers, for
840        example, SunCC recognizes above declarations as different functions and
841        generates broken code for hbpcre." I have changed the declarations to use
842        USPTR.
843    
844    16. GNU libtool is named differently on some systems. The autogen.sh script now
845        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
846        (FreeBSD).
847    
848    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
849        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
850        comment: "Figure out how to create a longlong from a string: strtoll and
851        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
852        instance, but it only takes 2 args instead of 3!"
853    
854    18. A subtle bug concerned with back references has been fixed by a change of
855        specification, with a corresponding code fix. A pattern such as
856        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
857        refers, was giving matches when it shouldn't. For example, xa=xaaa would
858        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
859        same bug. Such groups have to be quantified to be useful, or contained
860        inside another quantified group. (If there's no repetition, the reference
861        can never match.) The problem arises because, having left the group and
862        moved on to the rest of the pattern, a later failure that backtracks into
863        the group uses the captured value from the final iteration of the group
864        rather than the correct earlier one. I have fixed this in PCRE by forcing
865        any group that contains a reference to itself to be an atomic group; that
866        is, there cannot be any backtracking into it once it has completed. This is
867        similar to recursive and subroutine calls.
868    
869    
870    Version 8.00 19-Oct-09
871    ----------------------
872    
873    1.  The table for translating pcre_compile() error codes into POSIX error codes
874        was out-of-date, and there was no check on the pcre_compile() error code
875        being within the table. This could lead to an OK return being given in
876        error.
877    
878    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
879        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
880        in a Windows environment.
881    
882    3.  The pcregrep --count option prints the count for each file even when it is
883        zero, as does GNU grep. However, pcregrep was also printing all files when
884        --files-with-matches was added. Now, when both options are given, it prints
885        counts only for those files that have at least one match. (GNU grep just
886        prints the file name in this circumstance, but including the count seems
887        more useful - otherwise, why use --count?) Also ensured that the
888        combination -clh just lists non-zero counts, with no names.
889    
890    4.  The long form of the pcregrep -F option was incorrectly implemented as
891        --fixed_strings instead of --fixed-strings. This is an incompatible change,
892        but it seems right to fix it, and I didn't think it was worth preserving
893        the old behaviour.
894    
895    5.  The command line items --regex=pattern and --regexp=pattern were not
896        recognized by pcregrep, which required --regex pattern or --regexp pattern
897        (with a space rather than an '='). The man page documented the '=' forms,
898        which are compatible with GNU grep; these now work.
899    
900    6.  No libpcreposix.pc file was created for pkg-config; there was just
901        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
902    
903    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
904        when UCP support is not needed, by modifying the Python script that
905        generates it from Unicode data files. This should not matter if the module
906        is correctly used as a library, but I received one complaint about 50K of
907        unwanted data. My guess is that the person linked everything into his
908        program rather than using a library. Anyway, it does no harm.
909    
910    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
911        was a minimum greater than 1 for a wide character in a possessive
912        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
913        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
914        character. Chaos in the form of incorrect output or a compiling loop could
915        result.
916    
917    9.  The restrictions on what a pattern can contain when partial matching is
918        requested for pcre_exec() have been removed. All patterns can now be
919        partially matched by this function. In addition, if there are at least two
920        slots in the offset vector, the offset of the earliest inspected character
921        for the match and the offset of the end of the subject are set in them when
922        PCRE_ERROR_PARTIAL is returned.
923    
924    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
925        synonymous with PCRE_PARTIAL, for backwards compatibility, and
926        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
927        and may be more useful for multi-segment matching.
928    
929    11. Partial matching with pcre_exec() is now more intuitive. A partial match
930        used to be given if ever the end of the subject was reached; now it is
931        given only if matching could not proceed because another character was
932        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
933        string "Z", which now yields "no match" instead of "partial match". In the
934        case of pcre_dfa_exec(), "no match" is given if every matching path for the
935        final character ended with (*FAIL).
936    
937    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
938        if the pattern had a "must contain" character that was already found in the
939        earlier partial match, unless partial matching was again requested. For
940        example, with the pattern /dog.(body)?/, the "must contain" character is
941        "g". If the first part-match was for the string "dog", restarting with
942        "sbody" failed. This bug has been fixed.
943    
944    13. The string returned by pcre_dfa_exec() after a partial match has been
945        changed so that it starts at the first inspected character rather than the
946        first character of the match. This makes a difference only if the pattern
947        starts with a lookbehind assertion or \b or \B (\K is not supported by
948        pcre_dfa_exec()). It's an incompatible change, but it makes the two
949        matching functions compatible, and I think it's the right thing to do.
950    
951    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
952        so that the demonstration program is easily available in environments where
953        PCRE has not been installed from source.
954    
955    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
956        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
957        library.
958    
959    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
960        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
961        is not the first non-POSIX option to be added. Clearly some people find
962        these options useful.
963    
964    17. If a caller to the POSIX matching function regexec() passes a non-zero
965        value for nmatch with a NULL value for pmatch, the value of
966        nmatch is forced to zero.
967    
968    18. RunGrepTest did not have a test for the availability of the -u option of
969        the diff command, as RunTest does. It now checks in the same way as
970        RunTest, and also checks for the -b option.
971    
972    19. If an odd number of negated classes containing just a single character
973        interposed, within parentheses, between a forward reference to a named
974        subpattern and the definition of the subpattern, compilation crashed with
975        an internal error, complaining that it could not find the referenced
976        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
977        [The bug was that it was starting one character too far in when skipping
978        over the character class, thus treating the ] as data rather than
979        terminating the class. This meant it could skip too much.]
980    
981    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
982        /g option in pcretest when the pattern contains \K, which makes it possible
983        to have an empty string match not at the start, even when the pattern is
984        anchored. Updated pcretest and pcredemo to use this option.
985    
986    21. If the maximum number of capturing subpatterns in a recursion was greater
987        than the maximum at the outer level, the higher number was returned, but
988        with unset values at the outer level. The correct (outer level) value is
989        now given.
990    
991    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
992        PCRE did not set those parentheses (unlike Perl). I have now found a way to
993        make it do so. The string so far is captured, making this feature
994        compatible with Perl.
995    
996    23. The tests have been re-organized, adding tests 11 and 12, to make it
997        possible to check the Perl 5.10 features against Perl 5.10.
998    
999    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1000        pattern matches a fixed length string. PCRE did not allow this; now it
1001        does. Neither allows recursion.
1002    
1003    25. I finally figured out how to implement a request to provide the minimum
1004        length of subject string that was needed in order to match a given pattern.
1005        (It was back references and recursion that I had previously got hung up
1006        on.) This code has now been added to pcre_study(); it finds a lower bound
1007        to the length of subject needed. It is not necessarily the greatest lower
1008        bound, but using it to avoid searching strings that are too short does give
1009        some useful speed-ups. The value is available to calling programs via
1010        pcre_fullinfo().
1011    
1012    26. While implementing 25, I discovered to my embarrassment that pcretest had
1013        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1014        study optimizations had never been tested with that matching function.
1015        Oops. What is worse, even when it was passed study data, there was a bug in
1016        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1017        were also very few tests of studied patterns with pcre_dfa_exec().
1018    
1019    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1020        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1021        on the other side of the coin, they are no longer allowed to have different
1022        names, because these cannot be distinguished in PCRE, and this has caused
1023        confusion. (This is a difference from Perl.)
1024    
1025    28. When duplicate subpattern names are present (necessarily with different
1026        numbers, as required by 27 above), and a test is made by name in a
1027        conditional pattern, either for a subpattern having been matched, or for
1028        recursion in such a pattern, all the associated numbered subpatterns are
1029        tested, and the overall condition is true if the condition is true for any
1030        one of them. This is the way Perl works, and is also more like the way
1031        testing by number works.
1032    
1033    
1034    Version 7.9 11-Apr-09
1035    ---------------------
1036    
1037    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1038        (pcretest), all targets were linked against these libraries. This included
1039        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1040        libraries. This caused unwanted dependencies to be created. This problem
1041        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1042        pcretest is linked with readline.
1043    
1044    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1045        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1046        moved outside it again, because FALSE and TRUE are already defined in AIX,
1047        but BOOL is not.
1048    
1049    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1050        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1051    
1052    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1053        hyphens) following file names and line numbers when outputting matching
1054        lines. This is not true; no spaces are inserted. I have also clarified the
1055        wording for the --colour (or --color) option.
1056    
1057    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1058        was not coloured; this is different to GNU grep, so I have changed it to be
1059        the same.
1060    
1061    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1062        each matching line was coloured. Now it goes on to look for further matches
1063        of any of the test patterns, which is the same behaviour as GNU grep.
1064    
1065    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1066        doesn't make sense to accept an empty string match in pcregrep, so I have
1067        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1068        seems to be how GNU grep behaves.
1069    
1070    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1071        start or after a newline", because the conditional assertion was not being
1072        correctly handled. The rule now is that both the assertion and what follows
1073        in the first alternative must satisfy the test.
1074    
1075    9.  If auto-callout was enabled in a pattern with a conditional group whose
1076        condition was an assertion, PCRE could crash during matching, both with
1077        pcre_exec() and pcre_dfa_exec().
1078    
1079    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1080        used for matching.
1081    
1082    11. Unicode property support in character classes was not working for
1083        characters (bytes) greater than 127 when not in UTF-8 mode.
1084    
1085    12. Added the -M command line option to pcretest.
1086    
1087    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1088    
1089    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1090    
1091    16. Added comments and documentation about mis-use of no_arg in the C++
1092        wrapper.
1093    
1094    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1095        from Martin Jerabek that uses macro names for all relevant character and
1096        string constants.
1097    
1098    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1099        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1100        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1101        these, but not everybody uses configure.
1102    
1103    19. A conditional group that had only one branch was not being correctly
1104        recognized as an item that could match an empty string. This meant that an
1105        enclosing group might also not be so recognized, causing infinite looping
1106        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1107        with the subject "ab", where knowledge that the repeated group can match
1108        nothing is needed in order to break the loop.
1109    
1110    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1111        exec(), but without supplying a callout function, matching went wrong.
1112    
1113    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1114        leak if the size of the offset vector was greater than 30. When the vector
1115        is smaller, the saved offsets during recursion go onto a local stack
1116        vector, but for larger vectors malloc() is used. It was failing to free
1117        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1118        error, in fact).
1119    
1120    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1121        heapframe that is used only when UTF-8 support is enabled. This caused no
1122        problem, but was untidy.
1123    
1124    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1125        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1126        included within another project.
1127    
1128    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1129        slightly modified by me:
1130    
1131          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1132              not building pcregrep.
1133    
1134          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1135              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1136    
1137    25. Forward references, both numeric and by name, in patterns that made use of
1138        duplicate group numbers, could behave incorrectly or give incorrect errors,
1139        because when scanning forward to find the reference group, PCRE was not
1140        taking into account the duplicate group numbers. A pattern such as
1141        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1142    
1143    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1144        the feature of a custom pointer more persuasive (as requested by a user).
1145    
1146    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1147        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1148        pre-defined.
1149    
1150    28. Added support for (*UTF8) at the start of a pattern.
1151    
1152    29. Arrange for flags added by the "release type" setting in CMake to be shown
1153        in the configuration summary.
1154    
1155    
1156    Version 7.8 05-Sep-08
1157    ---------------------
1158    
1159    1.  Replaced UCP searching code with optimized version as implemented for Ad
1160        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1161        stage table and inline lookup instead of a function, giving speed ups of 2
1162        to 5 times on some simple patterns that I tested. Permission was given to
1163        distribute the MultiStage2.py script that generates the tables (it's not in
1164        the tarball, but is in the Subversion repository).
1165    
1166    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1167        scripts.
1168    
1169    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1170        a group with a zero qualifier. The result of the study could be incorrect,
1171        or the function might crash, depending on the pattern.
1172    
1173    4.  Caseless matching was not working for non-ASCII characters in back
1174        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1175        It now works when Unicode Property Support is available.
1176    
1177    5.  In pcretest, an escape such as \x{de} in the data was always generating
1178        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1179        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1180        truncation.
1181    
1182    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1183    
1184    7.  Added two (int) casts to pcregrep when printing the difference of two
1185        pointers, in case they are 64-bit values.
1186    
1187    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1188        test 2 if it fails.
1189    
1190    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1191        and a #define of that name to empty if it is not externally set. This is to
1192        allow users of MSVC to set it if necessary.
1193    
1194    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1195        the convenience functions in the pcre_get.c source file.
1196    
1197    11. An option change at the start of a pattern that had top-level alternatives
1198        could cause overwriting and/or a crash. This command provoked a crash in
1199        some environments:
1200    
1201          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1202    
1203        This potential security problem was recorded as CVE-2008-2371.
1204    
1205    12. For a pattern where the match had to start at the beginning or immediately
1206        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1207        pcre_dfa_exec() could read past the end of the passed subject if there was
1208        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1209        pcretest so that it places the subject at the end of its malloc-ed buffer.
1210    
1211    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1212        exec() might read past the end of the data buffer in UTF-8 mode.
1213    
1214    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1215        the data contained the byte 0x85 as part of a UTF-8 character within its
1216        first line. This applied both to normal and DFA matching.
1217    
1218    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1219        /^[^d]*?$/8 failed to match "abc".
1220    
1221    16. Added a missing copyright notice to pcrecpp_internal.h.
1222    
1223    17. Make it more clear in the documentation that values returned from
1224        pcre_exec() in ovector are byte offsets, not character counts.
1225    
1226    18. Tidied a few places to stop certain compilers from issuing warnings.
1227    
1228    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1229        supplied by Stefan Weber. I made a further small update for 7.8 because
1230        there is a change of source arrangements: the pcre_searchfuncs.c module is
1231        replaced by pcre_ucd.c.
1232    
1233    
1234    Version 7.7 07-May-08
1235    ---------------------
1236    
1237    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1238        a string to a long long, pretend we don't even have a long long." This is
1239        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1240    
1241    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1242        pre-7.6 versions, which defined a global no_arg variable instead of putting
1243        it in the RE class. (See also #8 below.)
1244    
1245    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1246        Lopes.
1247    
1248    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1249    
1250        (1) The include/exclude patterns were being applied to the whole pathnames
1251            of files, instead of just to the final components.
1252    
1253        (2) If there was more than one level of directory, the subdirectories were
1254            skipped unless they satisfied the include/exclude conditions. This is
1255            inconsistent with GNU grep (and could even be seen as contrary to the
1256            pcregrep specification - which I improved to make it absolutely clear).
1257            The action now is always to scan all levels of directory, and just
1258            apply the include/exclude patterns to regular files.
1259    
1260    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1261        --exclude_dir in the tests to avoid scanning .svn directories.
1262    
1263    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1264        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1265        doesn't support NULs in patterns.
1266    
1267    7.  Added some missing "const"s to declarations of static tables in
1268        pcre_compile.c and pcre_dfa_exec.c.
1269    
1270    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1271        caused by fix #2  above. (Subsequently also a second patch to fix the
1272        first patch. And a third patch - this was a messy problem.)
1273    
1274    9.  Applied Craig's patch to remove the use of push_back().
1275    
1276    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1277        matching function regexec().
1278    
1279    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1280        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1281        references. PCRE supports relative numbers with this syntax (I don't think
1282        Oniguruma does).
1283    
1284    12. Previously, a group with a zero repeat such as (...){0} was completely
1285        omitted from the compiled regex. However, this means that if the group
1286        was called as a subroutine from elsewhere in the pattern, things went wrong
1287        (an internal error was given). Such groups are now left in the compiled
1288        pattern, with a new opcode that causes them to be skipped at execution
1289        time.
1290    
1291    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1292        to the way PCRE behaves:
1293    
1294        (a) A lone ] character is dis-allowed (Perl treats it as data).
1295    
1296        (b) A back reference to an unmatched subpattern matches an empty string
1297            (Perl fails the current match path).
1298    
1299        (c) A data ] in a character class must be notated as \] because if the
1300            first data character in a class is ], it defines an empty class. (In
1301            Perl it is not possible to have an empty class.) The empty class []
1302            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1303            The negative empty class [^] matches any one character, independently
1304            of the DOTALL setting.
1305    
1306    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1307        non-existent subpattern following a character class starting with ']' and
1308        containing () gave an internal compiling error instead of "reference to
1309        non-existent subpattern". Fortunately, when the pattern did exist, the
1310        compiled code was correct. (When scanning forwards to check for the
1311        existencd of the subpattern, it was treating the data ']' as terminating
1312        the class, so got the count wrong. When actually compiling, the reference
1313        was subsequently set up correctly.)
1314    
1315    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1316        it was being rejected as not supported by pcre_dfa_exec(), even though
1317        other assertions are supported. I have made pcre_dfa_exec() support
1318        (*FAIL).
1319    
1320    16. The implementation of 13c above involved the invention of a new opcode,
1321        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1322        cannot be changed at match time, I realized I could make a small
1323        improvement to matching performance by compiling OP_ALLANY instead of
1324        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1325        on the OP_ANY path.
1326    
1327    17. Compiling pcretest on Windows with readline support failed without the
1328        following two fixes: (1) Make the unistd.h include conditional on
1329        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1330    
1331    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1332        ncurses library to be included for pcretest when ReadLine support is
1333        requested, but also to allow for it to be overridden. This patch came from
1334        Daniel Bergström.
1335    
1336    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1337        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1338        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1339        spotting this.
1340    
1341    
1342    Version 7.6 28-Jan-08
1343  ---------------------  ---------------------
1344    
1345   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one  1.  A character class containing a very large number of characters with
1346      that is more "standard", making use of automake and other autotools. There      codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1347        overflow.
1348    
1349    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1350        HAVE_LONG_LONG is not defined.
1351    
1352    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1353        bring them up to date and include new features. This patch includes:
1354    
1355        - Fixed PH's badly added libz and libbz2 support.
1356        - Fixed a problem with static linking.
1357        - Added pcredemo. [But later removed - see 7 below.]
1358        - Fixed dftables problem and added an option.
1359        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1360            HAVE_LONG_LONG.
1361        - Added readline support for pcretest.
1362        - Added an listing of the option settings after cmake has run.
1363    
1364    4.  A user submitted a patch to Makefile that makes it easy to create
1365        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1366        Makefile.am that cause it to include this special target, without
1367        affecting anything else. Note that the same mingw target plus all
1368        the other distribution libraries and programs are now supported
1369        when configuring with CMake (see 6 below) instead of with
1370        Configure/Make.
1371    
1372    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1373        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1374        exported in the Windows port". It has not yet been confirmed that the patch
1375        solves the problem, but it does no harm.
1376    
1377    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1378        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1379        with CMake, and also correct the comment about stack recursion.
1380    
1381    7.  Remove the automatic building of pcredemo from the ./configure system and
1382        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1383        of a program that users should build themselves after PCRE is installed, so
1384        building it automatically is not really right. What is more, it gave
1385        trouble in some build environments.
1386    
1387    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1388    
1389    
1390    Version 7.5 10-Jan-08
1391    ---------------------
1392    
1393    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1394        values in parens when parsing an RE using the C++ wrapper."
1395    
1396    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1397        Characters greater than 255 were excluded from the class instead of being
1398        included.
1399    
1400    3.  The same bug as (2) above applied to negated POSIX classes such as
1401        [:^space:].
1402    
1403    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1404        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1405        I have changed it.
1406    
1407    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1408        first named subpattern) and a construct such as (?&a) would reference the
1409        first named subpattern whose name started with "a" (in other words, the
1410        length check was missing). Both these problems are fixed. "Subpattern name
1411        expected" is now given for (?&) (a zero-length name), and this patch also
1412        makes it give the same error for \k'' (previously it complained that that
1413        was a reference to a non-existent subpattern).
1414    
1415    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1416        this is right because (?- can be followed by option settings as well as by
1417        digits. I have, however, made the messages clearer.
1418    
1419    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1420        than the number used in the conditional) now cause a compile-time error.
1421        This is actually not compatible with Perl, which accepts such patterns, but
1422        treats the conditional as always being FALSE (as PCRE used to), but it
1423        seems to me that giving a diagnostic is better.
1424    
1425    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1426        and messages.
1427    
1428    9.  Fix two occurrences of "backslash" in comments that should have been
1429        "backspace".
1430    
1431    10. Remove two redundant lines of code that can never be obeyed (their function
1432        was moved elsewhere).
1433    
1434    11. The program that makes PCRE's Unicode character property table had a bug
1435        which caused it to generate incorrect table entries for sequences of
1436        characters that have the same character type, but are in different scripts.
1437        It amalgamated them into a single range, with the script of the first of
1438        them. In other words, some characters were in the wrong script. There were
1439        thirteen such cases, affecting characters in the following ranges:
1440    
1441          U+002b0 - U+002c1
1442          U+0060c - U+0060d
1443          U+0061e - U+00612
1444          U+0064b - U+0065e
1445          U+0074d - U+0076d
1446          U+01800 - U+01805
1447          U+01d00 - U+01d77
1448          U+01d9b - U+01dbf
1449          U+0200b - U+0200f
1450          U+030fc - U+030fe
1451          U+03260 - U+0327f
1452          U+0fb46 - U+0fbb1
1453          U+10450 - U+1049d
1454    
1455    12. The -o option (show only the matching part of a line) for pcregrep was not
1456        compatible with GNU grep in that, if there was more than one match in a
1457        line, it showed only the first of them. It now behaves in the same way as
1458        GNU grep.
1459    
1460    13. If the -o and -v options were combined for pcregrep, it printed a blank
1461        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1462        does the same. The return code can be used to tell if there were any
1463        non-matching lines.
1464    
1465    14. Added --file-offsets and --line-offsets to pcregrep.
1466    
1467    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1468        infinitely looping recursion. The bug was that positive lookaheads were not
1469        being skipped when checking for a possible empty match (negative lookaheads
1470        and both kinds of lookbehind were skipped).
1471    
1472    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1473        inclusion of <windows.h> to before rather than after the definition of
1474        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1475    
1476    17. Specifying a possessive quantifier with a specific limit for a Unicode
1477        character property caused pcre_compile() to compile bad code, which led at
1478        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1479        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1480        caused the error; without that there was no problem.
1481    
1482    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1483    
1484    19. Added --enable-pcretest-libreadline.
1485    
1486    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1487        RE::GlobalReplace(). As a result, the number of replacements returned was
1488        double what it should be. I removed one of the increments, but Craig sent a
1489        later patch that removed the other one (the right fix) and added unit tests
1490        that check the return values (which was not done before).
1491    
1492    21. Several CMake things:
1493    
1494        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1495            the names libpcre and libpcreposix, not just pcre and pcreposix.
1496    
1497        (2) The above change means that pcretest and pcregrep are now correctly
1498            linked with the newly-built libraries, not previously installed ones.
1499    
1500        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1501    
1502    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1503        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1504        UTF-8 newline character). The key issue is that the pattern starts .*;
1505        this means that the match must be either at the beginning, or after a
1506        newline. The bug was in the code for advancing after a failed match and
1507        checking that the new position followed a newline. It was not taking
1508        account of UTF-8 characters correctly.
1509    
1510    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1511        character classes. PCRE was not treating the sequence [:...:] as a
1512        character class unless the ... were all letters. Perl, however, seems to
1513        allow any characters between [: and :], though of course it rejects as
1514        unknown any "names" that contain non-letters, because all the known class
1515        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1516        for example, whereas PCRE did not - it did not recognize a POSIX character
1517        class. This seemed a bit dangerous, so the code has been changed to be
1518        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1519        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1520        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1521        Perl does, and where it didn't before.
1522    
1523    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1524        Windows environments %n is disabled by default.
1525    
1526    
1527    Version 7.4 21-Sep-07
1528    ---------------------
1529    
1530    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1531        means that a class such as [\s] counted as "explicit reference to CR or
1532        LF". That isn't really right - the whole point of the change was to try to
1533        help when there was an actual mention of one of the two characters. So now
1534        the change happens only if \r or \n (or a literal CR or LF) character is
1535        encountered.
1536    
1537    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1538        of both had grown to the point where there were only 3 bits left.
1539        Fortunately, there was spare space in the data structure, and so I have
1540        moved the internal flags into a new 16-bit field to free up more option
1541        bits.
1542    
1543    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1544        but did not set the internal JCHANGED flag - either of these is enough to
1545        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1546        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1547        start sets both bits.
1548    
1549    4.  Added options (at build time, compile time, exec time) to change \R from
1550        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1551    
1552    5.  doc/pcresyntax.html was missing from the distribution.
1553    
1554    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1555        compatibility, even though it is no longer used.
1556    
1557    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1558        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1559        windows.h file is present (where different names are used). [This was
1560        reversed later after testing - see 16 below.]
1561    
1562    8.  Changed all #include <config.h> to #include "config.h". There were also
1563        some further <pcre.h> cases that I changed to "pcre.h".
1564    
1565    9.  When pcregrep was used with the --colour option, it missed the line ending
1566        sequence off the lines that it output.
1567    
1568    10. It was pointed out to me that arrays of string pointers cause lots of
1569        relocations when a shared library is dynamically loaded. A technique of
1570        using a single long string with a table of offsets can drastically reduce
1571        these. I have refactored PCRE in four places to do this. The result is
1572        dramatic:
1573    
1574          Originally:                          290
1575          After changing UCP table:            187
1576          After changing error message table:   43
1577          After changing table of "verbs"       36
1578          After changing table of Posix names   22
1579    
1580        Thanks to the folks working on Gregex for glib for this insight.
1581    
1582    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1583        unicode-properties was also set.
1584    
1585    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1586    
1587    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1588        checked only for CRLF.
1589    
1590    14. Added casts to pcretest.c to avoid compiler warnings.
1591    
1592    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1593    
1594    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1595        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1596        entirely. This removes changes made in 7 above.
1597    
1598    17. The CMake files have been updated, and there is now more information about
1599        building with CMake in the NON-UNIX-USE document.
1600    
1601    
1602    Version 7.3 28-Aug-07
1603    ---------------------
1604    
1605     1. In the rejigging of the build system that eventually resulted in 7.1, the
1606        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1607        brackets there is not right, since it causes compilers to look for an
1608        installed pcre.h, not the version that is in the source that is being
1609        compiled (which of course may be different). I have changed it back to:
1610    
1611          #include "pcre.h"
1612    
1613        I have a vague recollection that the change was concerned with compiling in
1614        different directories, but in the new build system, that is taken care of
1615        by the VPATH setting the Makefile.
1616    
1617     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1618        when the subject happened to end in the byte 0x85 (e.g. if the last
1619        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1620        characters but of course it shouldn't be taken as a newline when it is part
1621        of another character. The bug was that, for an unlimited repeat of . in
1622        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1623        characters when looking for a newline.
1624    
1625     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1626    
1627     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1628        in debug output.
1629    
1630     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1631        long printing in the pcrecpp unittest when running under MinGW.
1632    
1633     6. ESC_K was left out of the EBCDIC table.
1634    
1635     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1636        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1637        limit also applies to "virtual nesting" when a pattern is recursive, and in
1638        this case 1000 isn't so big. I have been able to remove this limit at the
1639        expense of backing off one optimization in certain circumstances. Normally,
1640        when pcre_exec() would call its internal match() function recursively and
1641        immediately return the result unconditionally, it uses a "tail recursion"
1642        feature to save stack. However, when a subpattern that can match an empty
1643        string has an unlimited repetition quantifier, it no longer makes this
1644        optimization. That gives it a stack frame in which to save the data for
1645        checking that an empty string has been matched. Previously this was taken
1646        from the 1000-entry workspace that had been reserved. So now there is no
1647        explicit limit, but more stack is used.
1648    
1649     8. Applied Daniel's patches to solve problems with the import/export magic
1650        syntax that is required for Windows, and which was going wrong for the
1651        pcreposix and pcrecpp parts of the library. These were overlooked when this
1652        problem was solved for the main library.
1653    
1654     9. There were some crude static tests to avoid integer overflow when computing
1655        the size of patterns that contain repeated groups with explicit upper
1656        limits. As the maximum quantifier is 65535, the maximum group length was
1657        set at 30,000 so that the product of these two numbers did not overflow a
1658        32-bit integer. However, it turns out that people want to use groups that
1659        are longer than 30,000 bytes (though not repeat them that many times).
1660        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1661        made it possible to implement the integer overflow checks in a much more
1662        dynamic way, which I have now done. The artificial limitation on group
1663        length has been removed - we now have only the limit on the total length of
1664        the compiled pattern, which depends on the LINK_SIZE setting.
1665    
1666    10. Fixed a bug in the documentation for get/copy named substring when
1667        duplicate names are permitted. If none of the named substrings are set, the
1668        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1669        empty string.
1670    
1671    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1672        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1673        because the ] is interpreted as the first data character and the
1674        terminating ] is not found. PCRE has been made compatible with Perl in this
1675        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1676        cause memory overwriting.
1677    
1678    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1679        string has been matched (to stop an infinite loop). It was not recognizing
1680        a conditional subpattern that could match an empty string if that
1681        subpattern was within another subpattern. For example, it looped when
1682        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1683        condition was not nested. This bug has been fixed.
1684    
1685    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1686        past the start of the subject in the presence of bytes with the top bit
1687        set, for example "\x8aBCD".
1688    
1689    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1690        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1691    
1692    14. Optimized (?!) to (*FAIL).
1693    
1694    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1695        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1696        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1697        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1698        does: it's just the validity check that is more restrictive.
1699    
1700    16. Inserted checks for integer overflows during escape sequence (backslash)
1701        processing, and also fixed erroneous offset values for syntax errors during
1702        backslash processing.
1703    
1704    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1705        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1706    
1707    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1708        caused an overrun.
1709    
1710    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1711        something other than just ASCII characters) inside a group that had an
1712        unlimited repeat caused a loop at compile time (while checking to see
1713        whether the group could match an empty string).
1714    
1715    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1716        [\P{Any}] did so. (Error in the code for printing property names.)
1717    
1718    21. An orphan \E inside a character class could cause a crash.
1719    
1720    22. A repeated capturing bracket such as (A)? could cause a wild memory
1721        reference during compilation.
1722    
1723    23. There are several functions in pcre_compile() that scan along a compiled
1724        expression for various reasons (e.g. to see if it's fixed length for look
1725        behind). There were bugs in these functions when a repeated \p or \P was
1726        present in the pattern. These operators have additional parameters compared
1727        with \d, etc, and these were not being taken into account when moving along
1728        the compiled data. Specifically:
1729    
1730        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1731            length.
1732    
1733        (b) An item such as \pL+ within a repeated group could cause crashes or
1734            loops.
1735    
1736        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1737            "reference to non-existent subpattern" error.
1738    
1739        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1740    
1741    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1742        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1743    
1744    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1745    
1746    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1747        character were causing crashes (broken optimization).
1748    
1749    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1750        \p or \P) caused a compile-time loop.
1751    
1752    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1753        break. For example, the unstudied pattern [\r\n]A does not match the string
1754        "\r\nA" because change 7.0/46 below moves the current point on by two
1755        characters after failing to match at the start. However, the pattern \nA
1756        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1757        the same is true. There doesn't seem any very clean way out of this, but
1758        what I have chosen to do makes the common cases work: PCRE now takes note
1759        of whether there can be an explicit match for \r or \n anywhere in the
1760        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1761        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1762        pattern has explicit CR or LF references.
1763    
1764    29. Added (*CR) etc for changing newline setting at start of pattern.
1765    
1766    
1767    Version 7.2 19-Jun-07
1768    ---------------------
1769    
1770     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1771        which is apparently normally available under Windows.
1772    
1773     2. Re-jig the pcregrep tests with different newline settings in an attempt
1774        to make them independent of the local environment's newline setting.
1775    
1776     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1777    
1778     4. Some of the "internals" tests were previously cut out when the link size
1779        was not 2, because the output contained actual offsets. The recent new
1780        "Z" feature of pcretest means that these can be cut out, making the tests
1781        usable with all link sizes.
1782    
1783     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1784        stack recursion. This gives a massive performance boost under BSD, but just
1785        a small improvement under Linux. However, it saves one field in the frame
1786        in all cases.
1787    
1788     6. Added more features from the forthcoming Perl 5.10:
1789    
1790        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1791            recursion call. It refers to the nth most recently opened parentheses.
1792    
1793        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1794            to be opened parentheses.
1795    
1796        (c) Conditions that refer to capturing parentheses can be specified
1797            relatively, for example, (?(-2)... or (?(+3)...
1798    
1799        (d) \K resets the start of the current match so that everything before
1800            is not part of it.
1801    
1802        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1803    
1804        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1805            reference syntax.
1806    
1807        (g) (?| introduces a group in which the numbering of parentheses in each
1808            alternative starts with the same number.
1809    
1810        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1811    
1812     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1813        PCRE_INFO_JCHANGED.
1814    
1815     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1816        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1817        for detecting groups that can match an empty string.
1818    
1819     9. A pattern with a very large number of alternatives (more than several
1820        hundred) was running out of internal workspace during the pre-compile
1821        phase, where pcre_compile() figures out how much memory will be needed. A
1822        bit of new cunning has reduced the workspace needed for groups with
1823        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1824        workspace instead of running out of the 4096 that are available.
1825    
1826    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1827    
1828    11. Applied patch from Google to remove an optimization that didn't quite work.
1829        The report of the bug said:
1830    
1831          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1832          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1833          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1834    
1835    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1836        it matched the wrong number of bytes.
1837    
1838    
1839    Version 7.1 24-Apr-07
1840    ---------------------
1841    
1842     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1843        that is more "standard", making use of automake and other Autotools. There
1844      is some re-arrangement of the files and adjustment of comments consequent      is some re-arrangement of the files and adjustment of comments consequent
1845      on this.      on this.
1846    
1847   2. I had an email from Eric Raymond about problems translating some of PCRE's   2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1848      man pages to HTML (despite the fact that I distribute HTML pages, some      for recursive directory scanning broke on some systems because the files
1849      people do their own conversions for various reasons). The problems      are not scanned in any specific order and on different systems the order
1850      concerned the use of low-level troff macros .br and .in. I have therefore      was different. A call to "sort" has been inserted into RunGrepTest for the
1851      removed all such uses from the man pages (some were redundant, some could      approprate test as a short-term fix. In the longer term there may be an
1852      be replaced by .nf/.fi pairs). The maintain/132html script that I use to      alternative.
1853      generate HTML has been updated to handle .nf/.fi and to complain if it  
1854      encounters .br or .in.   3. I had an email from Eric Raymond about problems translating some of PCRE's
1855        man pages to HTML (despite the fact that I distribute HTML pages, some
1856        people do their own conversions for various reasons). The problems
1857        concerned the use of low-level troff macros .br and .in. I have therefore
1858        removed all such uses from the man pages (some were redundant, some could
1859        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1860        HTML has been updated to handle .nf/.fi and to complain if it encounters
1861        .br or .in.
1862    
1863     4. Updated comments in configure.ac that get placed in config.h.in and also
1864        arranged for config.h to be included in the distribution, with the name
1865        config.h.generic, for the benefit of those who have to compile without
1866        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1867    
1868     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1869        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1870        makevp.bat for the current PCRE, using the additional files
1871        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1872    
1873     6. A Windows user reported a minor discrepancy with test 2, which turned out
1874        to be caused by a trailing space on an input line that had got lost in his
1875        copy. The trailing space was an accident, so I've just removed it.
1876    
1877     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1878        that is needed.
1879    
1880     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1881        as "const" (a) because they are and (b) because it helps the PHP
1882        maintainers who have recently made a script to detect big data structures
1883        in the php code that should be moved to the .rodata section. I remembered
1884        to update Builducptable as well, so it won't revert if ucptable.h is ever
1885        re-created.
1886    
1887     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1888        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1889        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1890        support is not required. This saves 1.5-2K of code, which is important in
1891        some applications.
1892    
1893        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1894        so as not to refer to the tables, even though these functions will never be
1895        called when UTF-8 support is disabled. Otherwise there are problems with a
1896        shared library.
1897    
1898    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1899    
1900        (a) It was defining its arguments as char * instead of void *.
1901    
1902        (b) It was assuming that all moves were upwards in memory; this was true
1903            a long time ago when I wrote it, but is no longer the case.
1904    
1905        The emulated memove() is provided for those environments that have neither
1906        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1907        is clearly not the case, as these two bugs were recently reported.
1908    
1909    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1910        and Detrail to create the HTML documentation, the .txt form of the man
1911        pages, and it removes trailing spaces from listed files. It also creates
1912        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1913        case, it wraps all the #defines with #ifndefs. This script should be run
1914        before "make dist".
1915    
1916    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1917        with Unicode property support.
1918    
1919        (a) For a maximizing quantifier, if the two different cases of the
1920            character were of different lengths in their UTF-8 codings (there are
1921            some cases like this - I found 11), and the matching function had to
1922            back up over a mixture of the two cases, it incorrectly assumed they
1923            were both the same length.
1924    
1925        (b) When PCRE was configured to use the heap rather than the stack for
1926            recursion during matching, it was not correctly preserving the data for
1927            the other case of a UTF-8 character when checking ahead for a match
1928            while processing a minimizing repeat. If the check also involved
1929            matching a wide character, but failed, corruption could cause an
1930            erroneous result when trying to check for a repeat of the original
1931            character.
1932    
1933    13. Some tidying changes to the testing mechanism:
1934    
1935        (a) The RunTest script now detects the internal link size and whether there
1936            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1937            values substituted by "configure". (The RunGrepTest script already did
1938            this for UTF-8.) The configure.ac script no longer substitutes the
1939            relevant variables.
1940    
1941        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1942            with length and offset values. This means that the output is different
1943            for different internal link sizes. Test 2 is skipped for link sizes
1944            other than 2 because of this, bypassing the problem. Unfortunately,
1945            there was also a test in test 3 (the locale tests) that used /B and
1946            failed for link sizes other than 2. Rather than cut the whole test out,
1947            I have added a new /Z option to pcretest that replaces the length and
1948            offset values with spaces. This is now used to make test 3 independent
1949            of link size. (Test 2 will be tidied up later.)
1950    
1951    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1952        segmentation fault instead of returning the appropriate error message.
1953    
1954    15. In multiline mode when the newline sequence was set to "any", the pattern
1955        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1956        This doesn't seem right; it now treats the CRLF combination as the line
1957        ending, and so does not match in that case. It's only a pattern such as ^$
1958        that would hit this one: something like ^ABC$ would have failed after \r
1959        and then tried again after \r\n.
1960    
1961    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1962        in an attempt to make files that differ only in their line terminators
1963        compare equal. This works on Linux.
1964    
1965    17. Under certain error circumstances pcregrep might try to free random memory
1966        as it exited. This is now fixed, thanks to valgrind.
1967    
1968    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1969        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1970        was because its rules for how to advance for /g after matching an empty
1971        string at the end of a line did not allow for this case. They now check for
1972        it specially.
1973    
1974    20. pcretest is supposed to handle patterns and data of any length, by
1975        extending its buffers when necessary. It was getting this wrong when the
1976        buffer for a data line had to be extended.
1977    
1978    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1979        CRLF as a newline sequence.
1980    
1981    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1982        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1983        I have nevertheless tidied it up.
1984    
1985    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1986    
1987    24. Added a man page for pcre-config.
1988    
1989    
1990  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06
1991  ---------------------  ---------------------

Legend:
Removed from v.99  
changed lines
  Added in v.753

  ViewVC Help
Powered by ViewVC 1.1.5