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

Diff of /code/trunk/ChangeLog

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

revision 327 by ph10, Sat Mar 8 19:38:30 2008 UTC revision 778 by ph10, Thu Dec 1 17:38:47 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.7 05-Mar-08  Version 8.21
5    ------------
6    
7    1.  Updating the JIT compiler.
8    
9    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10        are added as well.
11    
12    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14        calling _pcre_jit_exec. Some extra comments are added.
15    
16    4.  Mark settings inside atomic groups that do not contain any capturing
17        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18        was introduced by change 18 for 8.20.
19    
20    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21        ECMA-262 standard.
22    
23    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25        This bug was probably introduced by change 9 of 8.13.
26    
27    7.  While fixing 6 above, I noticed that a number of other items were being
28        incorrectly rejected as "not fixed length". This arose partly because newer
29        opcodes had not been added to the fixed-length checking code. I have (a)
30        corrected the bug and added tests for these items, and (b) arranged for an
31        error to occur if an unknown opcode is encountered while checking for fixed
32        length instead of just assuming "not fixed length". The items that were
33        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36    
37    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
38        being incorrectly compiled and would have given unpredicatble results.
39    
40    9.  A possessively repeated subpattern with minimum repeat count greater than
41        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
42        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
43        the first (A) could occur when it should not.
44    
45    10. Add a cast and remove a redundant test from the code.
46    
47    11. JIT should use pcre_malloc/pcre_free for allocation.
48    
49    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
50        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
51        is anything other than /usr, -L is still shown).
52    
53    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
54    
55    14. Perl does not support \N without a following name in a [] class; PCRE now
56        also gives an error.
57    
58    15. If a forward reference was repeated with an upper limit of around 2000,
59        it caused the error "internal error: overran compiling workspace". The
60        maximum number of forward references (including repeats) was limited by the
61        internal workspace, and dependent on the LINK_SIZE. The code has been
62        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
63        the default depends on LINK_SIZE. There is a new upper limit (for safety)
64        of around 200,000 forward references. While doing this, I also speeded up
65        the filling in of repeated forward references.
66    
67    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
68        incorrectly expecting the subject to contain another "a" after the start.
69    
70    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
71        in the match, the SKIP should be ignored. This was not happening; instead
72        the SKIP was being treated as NOMATCH. For patterns such as
73        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
74        tested.
75    
76    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
77        now much more compatible with Perl, in particular in cases where the result
78        is a non-match for a non-anchored pattern. For example, if
79        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
80        "m", where previously it did not return a name. A side effect of this
81        change is that for partial matches, the last encountered mark name is
82        returned, as for non matches. A number of tests that were previously not
83        Perl-compatible have been moved into the Perl-compatible test files. The
84        refactoring has had the pleasing side effect of removing one argument from
85        the match() function, thus reducing its stack requirements.
86    
87    19. If the /S+ option was used in pcretest to study a pattern using JIT,
88        subsequent uses of /S (without +) incorrectly behaved like /S+.
89    
90    21. Retrieve executable code size support for the JIT compiler and fixing
91        some warnings.
92    
93    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
94        not work when the shorter character appeared right at the end of the
95        subject string.
96    
97    
98    Version 8.20 21-Oct-2011
99    ------------------------
100    
101    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
102        a POSIX class. After further experiments with Perl, which convinced me that
103        Perl has bugs and confusions, a closing square bracket is no longer allowed
104        in a POSIX name. This bug also affected patterns with classes that started
105        with full stops.
106    
107    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
108        captured substring, but while checking the failing first alternative,
109        substring 1 is temporarily captured. If the output vector supplied to
110        pcre_exec() was not big enough for this capture, the yield of the function
111        was still zero ("insufficient space for captured substrings"). This cannot
112        be totally fixed without adding another stack variable, which seems a lot
113        of expense for a edge case. However, I have improved the situation in cases
114        such as /(a)(b)x|abc/ matched against "abc", where the return code
115        indicates that fewer than the maximum number of slots in the ovector have
116        been set.
117    
118    3.  Related to (2) above: when there are more back references in a pattern than
119        slots in the output vector, pcre_exec() uses temporary memory during
120        matching, and copies in the captures as far as possible afterwards. It was
121        using the entire output vector, but this conflicts with the specification
122        that only 2/3 is used for passing back captured substrings. Now it uses
123        only the first 2/3, for compatibility. This is, of course, another edge
124        case.
125    
126    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
127        main code base, and can be used by building with --enable-jit. When this is
128        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
129        runtime --no-jit option is given.
130    
131    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
132        ovector, the return from the function was zero, implying that there were
133        other matches that did not fit. The correct "exactly full" value is now
134        returned.
135    
136    6.  If a subpattern that was called recursively or as a subroutine contained
137        (*PRUNE) or any other control that caused it to give a non-standard return,
138        invalid errors such as "Error -26 (nested recursion at the same subject
139        position)" or even infinite loops could occur.
140    
141    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
142        computing the minimum length on reaching *ACCEPT, and so ended up with the
143        wrong value of 1 rather than 0. Further investigation indicates that
144        computing a minimum subject length in the presence of *ACCEPT is difficult
145        (think back references, subroutine calls), and so I have changed the code
146        so that no minimum is registered for a pattern that contains *ACCEPT.
147    
148    8.  If (*THEN) was present in the first (true) branch of a conditional group,
149        it was not handled as intended. [But see 16 below.]
150    
151    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
152        Sheri Pierce.
153    
154    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
155        the first byte in a match must be "a".
156    
157    11. Change 17 for 8.13 increased the recursion depth for patterns like
158        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
159        pattern contains any instances of (*THEN). If it does not, the old
160        optimizations are restored. It would be nice to do this on a per-group
161        basis, but at the moment that is not feasible.
162    
163    12. In some environments, the output of pcretest -C is CRLF terminated. This
164        broke RunTest's code that checks for the link size. A single white space
165        character after the value is now allowed for.
166    
167    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
168        For "fr", it uses the Windows-specific input and output files.
169    
170    14. If (*THEN) appeared in a group that was called recursively or as a
171        subroutine, it did not work as intended. [But see next item.]
172    
173    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
174        pattern fragments (but not containing any | characters). If A and B are
175        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
176        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
177        D. In other words, Perl considers parentheses that do not contain any |
178        characters to be part of a surrounding alternative, whereas PCRE was
179        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
180        differently. PCRE now behaves in the same way as Perl, except in the case
181        of subroutine/recursion calls such as (?1) which have in any case always
182        been different (but PCRE had them first :-).
183    
184    16. Related to 15 above: Perl does not treat the | in a conditional group as
185        creating alternatives. Such a group is treated in the same way as an
186        ordinary group without any | characters when processing (*THEN). PCRE has
187        been changed to match Perl's behaviour.
188    
189    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
190        RunGrepTest script failed.
191    
192    18. Change 22 for version 13 caused atomic groups to use more stack. This is
193        inevitable for groups that contain captures, but it can lead to a lot of
194        stack use in large patterns. The old behaviour has been restored for atomic
195        groups that do not contain any capturing parentheses.
196    
197    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
198        suppress the check for a minimum subject length at run time. (If it was
199        given to pcre_exec() or pcre_dfa_exec() it did work.)
200    
201    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
202        fail to work when decoding hex characters in data strings in EBCDIC
203        environments.
204    
205    21. It appears that in at least one Mac OS environment, the isxdigit() function
206        is implemented as a macro that evaluates to its argument more than once,
207        contravening the C 90 Standard (I haven't checked a later standard). There
208        was an instance in pcretest which caused it to go wrong when processing
209        \x{...} escapes in subject strings. The has been rewritten to avoid using
210        things like p++ in the argument of isxdigit().
211    
212    
213    Version 8.13 16-Aug-2011
214    ------------------------
215    
216    1.  The Unicode data tables have been updated to Unicode 6.0.0.
217    
218    2.  Two minor typos in pcre_internal.h have been fixed.
219    
220    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
221        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
222        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
223    
224    4.  There were a number of related bugs in the code for matching backrefences
225        caselessly in UTF-8 mode when codes for the characters concerned were
226        different numbers of bytes. For example, U+023A and U+2C65 are an upper
227        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
228        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
229        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
230        2-byte code at the end of the subject (it thought there wasn't enough data
231        left).
232    
233    5.  Comprehensive information about what went wrong is now returned by
234        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
235        as the output vector has at least 2 elements. The offset of the start of
236        the failing character and a reason code are placed in the vector.
237    
238    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
239        now returned is for the first byte of the failing character, instead of the
240        last byte inspected. This is an incompatible change, but I hope it is small
241        enough not to be a problem. It makes the returned offset consistent with
242        pcre_exec() and pcre_dfa_exec().
243    
244    7.  pcretest now gives a text phrase as well as the error number when
245        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
246        failure, the offset and reason code are output.
247    
248    8.  When \R was used with a maximizing quantifier it failed to skip backwards
249        over a \r\n pair if the subsequent match failed. Instead, it just skipped
250        back over a single character (\n). This seems wrong (because it treated the
251        two characters as a single entity when going forwards), conflicts with the
252        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
253        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
254        has been changed.
255    
256    9.  Some internal refactoring has changed the processing so that the handling
257        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
258        time (the PCRE_DOTALL option was changed this way some time ago: version
259        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
260        which was always a bit of a fudge. It also means that there is one less
261        argument for the match() function, which reduces its stack requirements
262        slightly. This change also fixes an incompatibility with Perl: the pattern
263        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
264    
265    10. More internal refactoring has drastically reduced the number of recursive
266        calls to match() for possessively repeated groups such as (abc)++ when
267        using pcre_exec().
268    
269    11. While implementing 10, a number of bugs in the handling of groups were
270        discovered and fixed:
271    
272        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
273        (a|)*(?1) gave a compile-time internal error.
274        ((a|)+)+  did not notice that the outer group could match an empty string.
275        (^a|^)+   was not marked as anchored.
276        (.*a|.*)+ was not marked as matching at start or after a newline.
277    
278    12. Yet more internal refactoring has removed another argument from the match()
279        function. Special calls to this function are now indicated by setting a
280        value in a variable in the "match data" data block.
281    
282    13. Be more explicit in pcre_study() instead of relying on "default" for
283        opcodes that mean there is no starting character; this means that when new
284        ones are added and accidentally left out of pcre_study(), testing should
285        pick them up.
286    
287    14. The -s option of pcretest has been documented for ages as being an old
288        synonym of -m (show memory usage). I have changed it to mean "force study
289        for every regex", that is, assume /S for every regex. This is similar to -i
290        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
291        using it. It makes it easier to run collections of tests with and without
292        study enabled, and thereby test pcre_study() more easily. All the standard
293        tests are now run with and without -s (but some patterns can be marked as
294        "never study" - see 20 below).
295    
296    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
297        restoration of the capturing data to the outer values was not happening
298        correctly.
299    
300    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
301        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
302        pattern had matched an empty string, and so incorrectly returned a no
303        match.
304    
305    17. There was optimizing code for the last branch of non-capturing parentheses,
306        and also for the obeyed branch of a conditional subexpression, which used
307        tail recursion to cut down on stack usage. Unfortunately, now that there is
308        the possibility of (*THEN) occurring in these branches, tail recursion is
309        no longer possible because the return has to be checked for (*THEN). These
310        two optimizations have therefore been removed. [But see 8.20/11 above.]
311    
312    18. If a pattern containing \R was studied, it was assumed that \R always
313        matched two bytes, thus causing the minimum subject length to be
314        incorrectly computed because \R can also match just one byte.
315    
316    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
317        was incorrectly computed.
318    
319    20. If /S is present twice on a test pattern in pcretest input, it now
320        *disables* studying, thereby overriding the use of -s on the command line
321        (see 14 above). This is necessary for one or two tests to keep the output
322        identical in both cases.
323    
324    21. When (*ACCEPT) was used in an assertion that matched an empty string and
325        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
326    
327    22. When an atomic group that contained a capturing parenthesis was
328        successfully matched, but the branch in which it appeared failed, the
329        capturing was not being forgotten if a higher numbered group was later
330        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
331        group 1 to "a", when in fact it should be unset. This applied to multi-
332        branched capturing and non-capturing groups, repeated or not, and also to
333        positive assertions (capturing in negative assertions does not happen
334        in PCRE) and also to nested atomic groups.
335    
336    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
337        subject after a captured substring, to make it easier to tell which of a
338        number of identical substrings has been captured.
339    
340    24. The way atomic groups are processed by pcre_exec() has been changed so that
341        if they are repeated, backtracking one repetition now resets captured
342        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
343        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
344        "aaa". Previously, it would have been "a". As part of this code
345        refactoring, the way recursive calls are handled has also been changed.
346    
347    25. If an assertion condition captured any substrings, they were not passed
348        back unless some other capturing happened later. For example, if
349        (?(?=(a))a) was matched against "a", no capturing was returned.
350    
351    26. When studying a pattern that contained subroutine calls or assertions,
352        the code for finding the minimum length of a possible match was handling
353        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
354        group 1 called group 2 while simultaneously a separate group 2 called group
355        1). A stack overflow occurred in this case. I have fixed this by limiting
356        the recursion depth to 10.
357    
358    27. Updated RunTest.bat in the distribution to the version supplied by Tom
359        Fortmann. This supports explicit test numbers on the command line, and has
360        argument validation and error reporting.
361    
362    28. An instance of \X with an unlimited repeat could fail if at any point the
363        first character it looked at was a mark character.
364    
365    29. Some minor code refactoring concerning Unicode properties and scripts
366        should reduce the stack requirement of match() slightly.
367    
368    30. Added the '=' option to pcretest to check the setting of unused capturing
369        slots at the end of the pattern, which are documented as being -1, but are
370        not included in the return count.
371    
372    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
373        compiled something random. Now it gives a compile-time error (as does
374        Perl).
375    
376    32. A *MARK encountered during the processing of a positive assertion is now
377        recorded and passed back (compatible with Perl).
378    
379    33. If --only-matching or --colour was set on a pcregrep call whose pattern
380        had alternative anchored branches, the search for a second match in a line
381        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
382        matched the line "0102" twice. The same bug affected patterns that started
383        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
384        twice.
385    
386    34. Previously, PCRE did not allow quantification of assertions. However, Perl
387        does, and because of capturing effects, quantifying parenthesized
388        assertions may at times be useful. Quantifiers are now allowed for
389        parenthesized assertions.
390    
391    35. A minor code tidy in pcre_compile() when checking options for \R usage.
392    
393    36. \g was being checked for fancy things in a character class, when it should
394        just be a literal "g".
395    
396    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
397        appearance of a nested POSIX class supersedes an apparent external class.
398        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
399        unescaped square brackets may also appear as part of class names. For
400        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
401        more like Perl. (But see 8.20/1 above.)
402    
403    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
404        was because it thought it was \N{name}, which is not supported).
405    
406    39. Add minix to OS list not supporting the -S option in pcretest.
407    
408    40. PCRE tries to detect cases of infinite recursion at compile time, but it
409        cannot analyze patterns in sufficient detail to catch mutual recursions
410        such as ((?1))((?2)). There is now a runtime test that gives an error if a
411        subgroup is called recursively as a subpattern for a second time at the
412        same position in the subject string. In previous releases this might have
413        been caught by the recursion limit, or it might have run out of stack.
414    
415    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
416        happen only once. PCRE was, however incorrectly giving a compile time error
417        "recursive call could loop indefinitely" because it cannot analyze the
418        pattern in sufficient detail. The compile time test no longer happens when
419        PCRE is compiling a conditional subpattern, but actual runaway loops are
420        now caught at runtime (see 40 above).
421    
422    42. It seems that Perl allows any characters other than a closing parenthesis
423        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
424        has been changed to be the same.
425    
426    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
427        as not to get warnings when autogen.sh is called. Also changed
428        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
429    
430    44. To help people who use pcregrep to scan files containing exceedingly long
431        lines, the following changes have been made:
432    
433        (a) The default value of the buffer size parameter has been increased from
434            8K to 20K. (The actual buffer used is three times this size.)
435    
436        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
437            PCRE is built.
438    
439        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
440            to be set at run time.
441    
442        (d) Numerical values in pcregrep options can be followed by K or M, for
443            example --buffer-size=50K.
444    
445        (e) If a line being scanned overflows pcregrep's buffer, an error is now
446            given and the return code is set to 2.
447    
448    45. Add a pointer to the latest mark to the callout data block.
449    
450    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
451        partial match of an empty string instead of no match. This was specific to
452        the use of ".".
453    
454    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
455        complete match instead of a partial match. This bug was dependent on both
456        the PCRE_UTF8 and PCRE_DOTALL options being set.
457    
458    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
459        starting byte set, because \b was not being ignored.
460    
461    
462    Version 8.12 15-Jan-2011
463    ------------------------
464    
465    1.  Fixed some typos in the markup of the man pages, and wrote a script that
466        checks for such things as part of the documentation building process.
467    
468    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
469        --match-limit and --recursion-limit options (added for 8.11). In
470        particular, this made one of the standard tests fail. (The integer value
471        went into the wrong half of a long int.)
472    
473    3.  If the --colour option was given to pcregrep with -v (invert match), it
474        did strange things, either producing crazy output, or crashing. It should,
475        of course, ignore a request for colour when reporting lines that do not
476        match.
477    
478    4.  Another pcregrep bug caused similar problems if --colour was specified with
479        -M (multiline) and the pattern match finished with a line ending.
480    
481    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
482        matched in multiline mode, the following line was shown as part of the
483        match. This seems wrong, so I have changed it.
484    
485    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
486        the check for further matches in the same line (so they could be coloured)
487        to overrun the end of the current line. If another match was found, it was
488        incorrectly shown (and then shown again when found in the next line).
489    
490    7.  If pcregrep was compiled under Windows, there was a reference to the
491        function pcregrep_exit() before it was defined. I am assuming this was
492        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
493        reported by a user. I've moved the definition above the reference.
494    
495    
496    Version 8.11 10-Dec-2010
497    ------------------------
498    
499    1.  (*THEN) was not working properly if there were untried alternatives prior
500        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
501        backtracked to try for "b" instead of moving to the next alternative branch
502        at the same level (in this case, to look for "c"). The Perl documentation
503        is clear that when (*THEN) is backtracked onto, it goes to the "next
504        alternative in the innermost enclosing group".
505    
506    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
507        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
508        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
509        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
510        (*THEN).
511    
512    3.  If \s appeared in a character class, it removed the VT character from
513        the class, even if it had been included by some previous item, for example
514        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
515        of \s, but is part of the POSIX "space" class.)
516    
517    4.  A partial match never returns an empty string (because you can always
518        match an empty string at the end of the subject); however the checking for
519        an empty string was starting at the "start of match" point. This has been
520        changed to the "earliest inspected character" point, because the returned
521        data for a partial match starts at this character. This means that, for
522        example, /(?<=abc)def/ gives a partial match for the subject "abc"
523        (previously it gave "no match").
524    
525    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
526        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
527        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
528        has an implication that the given string is incomplete (because a partial
529        match is preferred over a full match). For this reason, these items now
530        give a partial match in this situation. [Aside: previously, the one case
531        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
532        match rather than a full match, which was wrong by the old rules, but is
533        now correct.]
534    
535    6.  There was a bug in the handling of #-introduced comments, recognized when
536        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
537        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
538        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
539        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
540        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
541        places in pcre_compile().
542    
543    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
544        comments when looking ahead for named forward references to subpatterns,
545        the only newline sequence it recognized was NL. It now handles newlines
546        according to the set newline convention.
547    
548    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
549        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
550        cater for a lack of strerror(). These oversights have been fixed.
551    
552    9.  Added --match-limit and --recursion-limit to pcregrep.
553    
554    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
555    
556    11. When the -o option was used, pcregrep was setting a return code of 1, even
557        when matches were found, and --line-buffered was not being honoured.
558    
559    12. Added an optional parentheses number to the -o and --only-matching options
560        of pcregrep.
561    
562    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
563        can match an empty string. The code to do it in pcretest and pcredemo
564        needed fixing:
565    
566        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
567            only one byte after an empty string match just before CRLF (this case
568            just got forgotten; "any" and "anycrlf" were OK).
569    
570        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
571            mode when an empty string match preceded an ASCII character followed by
572            a non-ASCII character. (The code for advancing by one character rather
573            than one byte was nonsense.)
574    
575        (c) The pcredemo.c sample program did not have any code at all to handle
576            the cases when CRLF is a valid newline sequence.
577    
578    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
579        as a starting offset was within the subject string. There is now a new
580        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
581        negative or greater than the length of the string. In order to test this,
582        pcretest is extended to allow the setting of negative starting offsets.
583    
584    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
585        starting offset points to the beginning of a UTF-8 character was
586        unnecessarily clumsy. I tidied it up.
587    
588    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
589        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
590    
591    17. Nobody had reported that the --include_dir option, which was added in
592        release 7.7 should have been called --include-dir (hyphen, not underscore)
593        for compatibility with GNU grep. I have changed it to --include-dir, but
594        left --include_dir as an undocumented synonym, and the same for
595        --exclude-dir, though that is not available in GNU grep, at least as of
596        release 2.5.4.
597    
598    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
599        characters from a string of bytes) have been redefined so as not to use
600        loops, in order to improve performance in some environments. At the same
601        time, I abstracted some of the common code into auxiliary macros to save
602        repetition (this should not affect the compiled code).
603    
604    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
605        compile-time error is now given if \c is not followed by an ASCII
606        character, that is, a byte less than 128. (In EBCDIC mode, the code is
607        different, and any byte value is allowed.)
608    
609    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
610        START_OPTIMIZE option, which is now allowed at compile time - but just
611        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
612        to pcregrep and other applications that have no direct access to PCRE
613        options. The new /Y option in pcretest sets this option when calling
614        pcre_compile().
615    
616    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
617        back references. Groups containing recursive back references were forced to
618        be atomic by that change, but in the case of named groups, the amount of
619        memory required was incorrectly computed, leading to "Failed: internal
620        error: code overflow". This has been fixed.
621    
622    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
623        pcretest.c, to avoid build problems in some Borland environments.
624    
625    
626    Version 8.10 25-Jun-2010
627    ------------------------
628    
629    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
630        THEN.
631    
632    2.  (*ACCEPT) was not working when inside an atomic group.
633    
634    3.  Inside a character class, \B is treated as a literal by default, but
635        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
636        causes the error). The code is unchanged, but I tidied the documentation.
637    
638    4.  Inside a character class, PCRE always treated \R and \X as literals,
639        whereas Perl faults them if its -w option is set. I have changed PCRE so
640        that it faults them when PCRE_EXTRA is set.
641    
642    5.  Added support for \N, which always matches any character other than
643        newline. (It is the same as "." when PCRE_DOTALL is not set.)
644    
645    6.  When compiling pcregrep with newer versions of gcc which may have
646        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
647        declared with attribute warn_unused_result" were given. Just casting the
648        result to (void) does not stop the warnings; a more elaborate fudge is
649        needed. I've used a macro to implement this.
650    
651    7.  Minor change to pcretest.c to avoid a compiler warning.
652    
653    8.  Added four artifical Unicode properties to help with an option to make
654        \s etc use properties (see next item). The new properties are: Xan
655        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
656    
657    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
658        use Unicode properties. (*UCP) at the start of a pattern can be used to set
659        this option. Modified pcretest to add /W to test this facility. Added
660        REG_UCP to make it available via the POSIX interface.
661    
662    10. Added --line-buffered to pcregrep.
663    
664    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
665        studied, and the match started with a letter with a code point greater than
666        127 whose first byte was different to the first byte of the other case of
667        the letter, the other case of this starting letter was not recognized
668        (#976).
669    
670    12. If a pattern that was studied started with a repeated Unicode property
671        test, for example, \p{Nd}+, there was the theoretical possibility of
672        setting up an incorrect bitmap of starting bytes, but fortunately it could
673        not have actually happened in practice until change 8 above was made (it
674        added property types that matched character-matching opcodes).
675    
676    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
677        possible starting bytes for non-anchored patterns.
678    
679    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
680        \R, and also a number of cases that involve Unicode properties, both
681        explicit and implicit when PCRE_UCP is set.
682    
683    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
684        input, it could crash or give wrong results if characters with values
685        greater than 0xc0 were present in the subject string. (Detail: it assumed
686        UTF-8 input when processing these items.)
687    
688    16. Added a lot of (int) casts to avoid compiler warnings in systems where
689        size_t is 64-bit (#991).
690    
691    17. Added a check for running out of memory when PCRE is compiled with
692        --disable-stack-for-recursion (#990).
693    
694    18. If the last data line in a file for pcretest does not have a newline on
695        the end, a newline was missing in the output.
696    
697    19. The default pcre_chartables.c file recognizes only ASCII characters (values
698        less than 128) in its various bitmaps. However, there is a facility for
699        generating tables according to the current locale when PCRE is compiled. It
700        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
701        characters, are recognized by isspace() and therefore were getting set in
702        these tables, and indeed these tables seem to approximate to ISO 8859. This
703        caused a problem in UTF-8 mode when pcre_study() was used to create a list
704        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
705        which of course cannot start UTF-8 characters. I have changed the code so
706        that only real ASCII characters (less than 128) and the correct starting
707        bytes for UTF-8 encodings are set for characters greater than 127 when in
708        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
709        altogether.)
710    
711    20. Added the /T option to pcretest so as to be able to run tests with non-
712        standard character tables, thus making it possible to include the tests
713        used for 19 above in the standard set of tests.
714    
715    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
716        reference to a subpattern the other side of a comment that contains an
717        opening parenthesis caused either an internal compiling error, or a
718        reference to the wrong subpattern.
719    
720    
721    Version 8.02 19-Mar-2010
722    ------------------------
723    
724    1.  The Unicode data tables have been updated to Unicode 5.2.0.
725    
726    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
727        configured.
728    
729    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
730        original author of that file, following a query about its status.
731    
732    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
733        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
734    
735    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
736        quantifier applied to a forward-referencing subroutine call, could compile
737        incorrect code or give the error "internal error: previously-checked
738        referenced subpattern not found".
739    
740    6.  Both MS Visual Studio and Symbian OS have problems with initializing
741        variables to point to external functions. For these systems, therefore,
742        pcre_malloc etc. are now initialized to local functions that call the
743        relevant global functions.
744    
745    7.  There were two entries missing in the vectors called coptable and poptable
746        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
747        I've fixed the data, and added a kludgy way of testing at compile time that
748        the lengths are correct (equal to the number of opcodes).
749    
750    8.  Following on from 7, I added a similar kludge to check the length of the
751        eint vector in pcreposix.c.
752    
753    9.  Error texts for pcre_compile() are held as one long string to avoid too
754        much relocation at load time. To find a text, the string is searched,
755        counting zeros. There was no check for running off the end of the string,
756        which could happen if a new error number was added without updating the
757        string.
758    
759    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
760    
761    11. \K was not working if it appeared in an atomic group or in a group that
762        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
763        \K is "not well defined" if used in an assertion. PCRE now accepts it if
764        the assertion is positive, but not if it is negative.
765    
766    12. Change 11 fortuitously reduced the size of the stack frame used in the
767        "match()" function of pcre_exec.c by one pointer. Forthcoming
768        implementation of support for (*MARK) will need an extra pointer on the
769        stack; I have reserved it now, so that the stack frame size does not
770        decrease.
771    
772    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
773        item in branch that calls a recursion is a subroutine call - as in the
774        second branch in the above example - was incorrectly given the compile-
775        time error "recursive call could loop indefinitely" because pcre_compile()
776        was not correctly checking the subroutine for matching a non-empty string.
777    
778    14. The checks for overrunning compiling workspace could trigger after an
779        overrun had occurred. This is a "should never occur" error, but it can be
780        triggered by pathological patterns such as hundreds of nested parentheses.
781        The checks now trigger 100 bytes before the end of the workspace.
782    
783    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
784    
785    
786    Version 8.01 19-Jan-2010
787    ------------------------
788    
789    1.  If a pattern contained a conditional subpattern with only one branch (in
790        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
791        computed the wrong minimum data length (which is of course zero for such
792        subpatterns). This could cause incorrect "no match" results.
793    
794    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
795        the pattern is reset in the first branch, pcre_compile() failed with
796        "internal error: code overflow at offset...". This happened only when
797        the reset was to the original external option setting. (An optimization
798        abstracts leading options settings into an external setting, which was the
799        cause of this.)
800    
801    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
802        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
803        assertion pattern did not match (meaning that the assertion was true), it
804        was incorrectly treated as false if the SKIP had been reached during the
805        matching. This also applied to assertions used as conditions.
806    
807    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
808        assertion subpattern, including such a pattern used as a condition,
809        unpredictable results occurred, instead of the error return
810        PCRE_ERROR_DFA_UITEM.
811    
812    5.  The C++ GlobalReplace function was not working like Perl for the special
813        situation when an empty string is matched. It now does the fancy magic
814        stuff that is necessary.
815    
816    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
817        removed. (These were left over from very, very early versions of PCRE.)
818    
819    7.  Some cosmetic changes to the code to make life easier when compiling it
820        as part of something else:
821    
822        (a) Change DEBUG to PCRE_DEBUG.
823    
824        (b) In pcre_compile(), rename the member of the "branch_chain" structure
825            called "current" as "current_branch", to prevent a collision with the
826            Linux macro when compiled as a kernel module.
827    
828        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
829            prevent a collision with the Linux macro when compiled as a kernel
830            module.
831    
832    8.  In pcre_compile() there are some checks for integer overflows that used to
833        cast potentially large values to (double). This has been changed to that
834        when building, a check for int64_t is made, and if it is found, it is used
835        instead, thus avoiding the use of floating point arithmetic. (There is no
836        other use of FP in PCRE.) If int64_t is not found, the fallback is to
837        double.
838    
839    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
840        2005 (difference between two addresses compared to an unsigned value).
841    
842    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
843        custom one, because of the following reported problem in Windows:
844    
845          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
846              under Win32.
847          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
848              therefore missing the function definition.
849          - The compiler thus generates a "C" signature for the test function.
850          - The linker fails to find the "C" function.
851          - PCRE fails to configure if asked to do so against libbz2.
852    
853    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
854        messages were output:
855    
856          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
857          rerunning libtoolize, to keep the correct libtool macros in-tree.
858          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
859    
860        I have done both of these things.
861    
862    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
863        most of the time, it *can* run out if it is given a pattern that contains a
864        runaway infinite recursion. I updated the discussion in the pcrestack man
865        page.
866    
867    13. Now that we have gone to the x.xx style of version numbers, the minor
868        version may start with zero. Using 08 or 09 is a bad idea because users
869        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
870        interpreted as invalid octal numbers. I've updated the previous comment in
871        configure.ac, and also added a check that gives an error if 08 or 09 are
872        used.
873    
874    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
875        causing partial matching to fail when the end of the subject matched \W
876        in a UTF-8 pattern where \W was quantified with a minimum of 3.
877    
878    15. There were some discrepancies between the declarations in pcre_internal.h
879        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
880        their definitions. The declarations used "const uschar *" and the
881        definitions used USPTR. Even though USPTR is normally defined as "const
882        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
883        reported that: "This difference in casting confuses some C++ compilers, for
884        example, SunCC recognizes above declarations as different functions and
885        generates broken code for hbpcre." I have changed the declarations to use
886        USPTR.
887    
888    16. GNU libtool is named differently on some systems. The autogen.sh script now
889        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
890        (FreeBSD).
891    
892    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
893        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
894        comment: "Figure out how to create a longlong from a string: strtoll and
895        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
896        instance, but it only takes 2 args instead of 3!"
897    
898    18. A subtle bug concerned with back references has been fixed by a change of
899        specification, with a corresponding code fix. A pattern such as
900        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
901        refers, was giving matches when it shouldn't. For example, xa=xaaa would
902        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
903        same bug. Such groups have to be quantified to be useful, or contained
904        inside another quantified group. (If there's no repetition, the reference
905        can never match.) The problem arises because, having left the group and
906        moved on to the rest of the pattern, a later failure that backtracks into
907        the group uses the captured value from the final iteration of the group
908        rather than the correct earlier one. I have fixed this in PCRE by forcing
909        any group that contains a reference to itself to be an atomic group; that
910        is, there cannot be any backtracking into it once it has completed. This is
911        similar to recursive and subroutine calls.
912    
913    
914    Version 8.00 19-Oct-09
915    ----------------------
916    
917    1.  The table for translating pcre_compile() error codes into POSIX error codes
918        was out-of-date, and there was no check on the pcre_compile() error code
919        being within the table. This could lead to an OK return being given in
920        error.
921    
922    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
923        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
924        in a Windows environment.
925    
926    3.  The pcregrep --count option prints the count for each file even when it is
927        zero, as does GNU grep. However, pcregrep was also printing all files when
928        --files-with-matches was added. Now, when both options are given, it prints
929        counts only for those files that have at least one match. (GNU grep just
930        prints the file name in this circumstance, but including the count seems
931        more useful - otherwise, why use --count?) Also ensured that the
932        combination -clh just lists non-zero counts, with no names.
933    
934    4.  The long form of the pcregrep -F option was incorrectly implemented as
935        --fixed_strings instead of --fixed-strings. This is an incompatible change,
936        but it seems right to fix it, and I didn't think it was worth preserving
937        the old behaviour.
938    
939    5.  The command line items --regex=pattern and --regexp=pattern were not
940        recognized by pcregrep, which required --regex pattern or --regexp pattern
941        (with a space rather than an '='). The man page documented the '=' forms,
942        which are compatible with GNU grep; these now work.
943    
944    6.  No libpcreposix.pc file was created for pkg-config; there was just
945        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
946    
947    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
948        when UCP support is not needed, by modifying the Python script that
949        generates it from Unicode data files. This should not matter if the module
950        is correctly used as a library, but I received one complaint about 50K of
951        unwanted data. My guess is that the person linked everything into his
952        program rather than using a library. Anyway, it does no harm.
953    
954    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
955        was a minimum greater than 1 for a wide character in a possessive
956        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
957        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
958        character. Chaos in the form of incorrect output or a compiling loop could
959        result.
960    
961    9.  The restrictions on what a pattern can contain when partial matching is
962        requested for pcre_exec() have been removed. All patterns can now be
963        partially matched by this function. In addition, if there are at least two
964        slots in the offset vector, the offset of the earliest inspected character
965        for the match and the offset of the end of the subject are set in them when
966        PCRE_ERROR_PARTIAL is returned.
967    
968    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
969        synonymous with PCRE_PARTIAL, for backwards compatibility, and
970        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
971        and may be more useful for multi-segment matching.
972    
973    11. Partial matching with pcre_exec() is now more intuitive. A partial match
974        used to be given if ever the end of the subject was reached; now it is
975        given only if matching could not proceed because another character was
976        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
977        string "Z", which now yields "no match" instead of "partial match". In the
978        case of pcre_dfa_exec(), "no match" is given if every matching path for the
979        final character ended with (*FAIL).
980    
981    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
982        if the pattern had a "must contain" character that was already found in the
983        earlier partial match, unless partial matching was again requested. For
984        example, with the pattern /dog.(body)?/, the "must contain" character is
985        "g". If the first part-match was for the string "dog", restarting with
986        "sbody" failed. This bug has been fixed.
987    
988    13. The string returned by pcre_dfa_exec() after a partial match has been
989        changed so that it starts at the first inspected character rather than the
990        first character of the match. This makes a difference only if the pattern
991        starts with a lookbehind assertion or \b or \B (\K is not supported by
992        pcre_dfa_exec()). It's an incompatible change, but it makes the two
993        matching functions compatible, and I think it's the right thing to do.
994    
995    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
996        so that the demonstration program is easily available in environments where
997        PCRE has not been installed from source.
998    
999    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1000        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1001        library.
1002    
1003    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1004        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1005        is not the first non-POSIX option to be added. Clearly some people find
1006        these options useful.
1007    
1008    17. If a caller to the POSIX matching function regexec() passes a non-zero
1009        value for nmatch with a NULL value for pmatch, the value of
1010        nmatch is forced to zero.
1011    
1012    18. RunGrepTest did not have a test for the availability of the -u option of
1013        the diff command, as RunTest does. It now checks in the same way as
1014        RunTest, and also checks for the -b option.
1015    
1016    19. If an odd number of negated classes containing just a single character
1017        interposed, within parentheses, between a forward reference to a named
1018        subpattern and the definition of the subpattern, compilation crashed with
1019        an internal error, complaining that it could not find the referenced
1020        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1021        [The bug was that it was starting one character too far in when skipping
1022        over the character class, thus treating the ] as data rather than
1023        terminating the class. This meant it could skip too much.]
1024    
1025    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1026        /g option in pcretest when the pattern contains \K, which makes it possible
1027        to have an empty string match not at the start, even when the pattern is
1028        anchored. Updated pcretest and pcredemo to use this option.
1029    
1030    21. If the maximum number of capturing subpatterns in a recursion was greater
1031        than the maximum at the outer level, the higher number was returned, but
1032        with unset values at the outer level. The correct (outer level) value is
1033        now given.
1034    
1035    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1036        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1037        make it do so. The string so far is captured, making this feature
1038        compatible with Perl.
1039    
1040    23. The tests have been re-organized, adding tests 11 and 12, to make it
1041        possible to check the Perl 5.10 features against Perl 5.10.
1042    
1043    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1044        pattern matches a fixed length string. PCRE did not allow this; now it
1045        does. Neither allows recursion.
1046    
1047    25. I finally figured out how to implement a request to provide the minimum
1048        length of subject string that was needed in order to match a given pattern.
1049        (It was back references and recursion that I had previously got hung up
1050        on.) This code has now been added to pcre_study(); it finds a lower bound
1051        to the length of subject needed. It is not necessarily the greatest lower
1052        bound, but using it to avoid searching strings that are too short does give
1053        some useful speed-ups. The value is available to calling programs via
1054        pcre_fullinfo().
1055    
1056    26. While implementing 25, I discovered to my embarrassment that pcretest had
1057        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1058        study optimizations had never been tested with that matching function.
1059        Oops. What is worse, even when it was passed study data, there was a bug in
1060        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1061        were also very few tests of studied patterns with pcre_dfa_exec().
1062    
1063    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1064        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1065        on the other side of the coin, they are no longer allowed to have different
1066        names, because these cannot be distinguished in PCRE, and this has caused
1067        confusion. (This is a difference from Perl.)
1068    
1069    28. When duplicate subpattern names are present (necessarily with different
1070        numbers, as required by 27 above), and a test is made by name in a
1071        conditional pattern, either for a subpattern having been matched, or for
1072        recursion in such a pattern, all the associated numbered subpatterns are
1073        tested, and the overall condition is true if the condition is true for any
1074        one of them. This is the way Perl works, and is also more like the way
1075        testing by number works.
1076    
1077    
1078    Version 7.9 11-Apr-09
1079    ---------------------
1080    
1081    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1082        (pcretest), all targets were linked against these libraries. This included
1083        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1084        libraries. This caused unwanted dependencies to be created. This problem
1085        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1086        pcretest is linked with readline.
1087    
1088    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1089        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1090        moved outside it again, because FALSE and TRUE are already defined in AIX,
1091        but BOOL is not.
1092    
1093    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1094        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1095    
1096    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1097        hyphens) following file names and line numbers when outputting matching
1098        lines. This is not true; no spaces are inserted. I have also clarified the
1099        wording for the --colour (or --color) option.
1100    
1101    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1102        was not coloured; this is different to GNU grep, so I have changed it to be
1103        the same.
1104    
1105    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1106        each matching line was coloured. Now it goes on to look for further matches
1107        of any of the test patterns, which is the same behaviour as GNU grep.
1108    
1109    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1110        doesn't make sense to accept an empty string match in pcregrep, so I have
1111        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1112        seems to be how GNU grep behaves.
1113    
1114    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1115        start or after a newline", because the conditional assertion was not being
1116        correctly handled. The rule now is that both the assertion and what follows
1117        in the first alternative must satisfy the test.
1118    
1119    9.  If auto-callout was enabled in a pattern with a conditional group whose
1120        condition was an assertion, PCRE could crash during matching, both with
1121        pcre_exec() and pcre_dfa_exec().
1122    
1123    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1124        used for matching.
1125    
1126    11. Unicode property support in character classes was not working for
1127        characters (bytes) greater than 127 when not in UTF-8 mode.
1128    
1129    12. Added the -M command line option to pcretest.
1130    
1131    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1132    
1133    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1134    
1135    16. Added comments and documentation about mis-use of no_arg in the C++
1136        wrapper.
1137    
1138    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1139        from Martin Jerabek that uses macro names for all relevant character and
1140        string constants.
1141    
1142    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1143        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1144        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1145        these, but not everybody uses configure.
1146    
1147    19. A conditional group that had only one branch was not being correctly
1148        recognized as an item that could match an empty string. This meant that an
1149        enclosing group might also not be so recognized, causing infinite looping
1150        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1151        with the subject "ab", where knowledge that the repeated group can match
1152        nothing is needed in order to break the loop.
1153    
1154    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1155        exec(), but without supplying a callout function, matching went wrong.
1156    
1157    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1158        leak if the size of the offset vector was greater than 30. When the vector
1159        is smaller, the saved offsets during recursion go onto a local stack
1160        vector, but for larger vectors malloc() is used. It was failing to free
1161        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1162        error, in fact).
1163    
1164    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1165        heapframe that is used only when UTF-8 support is enabled. This caused no
1166        problem, but was untidy.
1167    
1168    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1169        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1170        included within another project.
1171    
1172    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1173        slightly modified by me:
1174    
1175          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1176              not building pcregrep.
1177    
1178          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1179              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1180    
1181    25. Forward references, both numeric and by name, in patterns that made use of
1182        duplicate group numbers, could behave incorrectly or give incorrect errors,
1183        because when scanning forward to find the reference group, PCRE was not
1184        taking into account the duplicate group numbers. A pattern such as
1185        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1186    
1187    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1188        the feature of a custom pointer more persuasive (as requested by a user).
1189    
1190    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1191        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1192        pre-defined.
1193    
1194    28. Added support for (*UTF8) at the start of a pattern.
1195    
1196    29. Arrange for flags added by the "release type" setting in CMake to be shown
1197        in the configuration summary.
1198    
1199    
1200    Version 7.8 05-Sep-08
1201    ---------------------
1202    
1203    1.  Replaced UCP searching code with optimized version as implemented for Ad
1204        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1205        stage table and inline lookup instead of a function, giving speed ups of 2
1206        to 5 times on some simple patterns that I tested. Permission was given to
1207        distribute the MultiStage2.py script that generates the tables (it's not in
1208        the tarball, but is in the Subversion repository).
1209    
1210    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1211        scripts.
1212    
1213    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1214        a group with a zero qualifier. The result of the study could be incorrect,
1215        or the function might crash, depending on the pattern.
1216    
1217    4.  Caseless matching was not working for non-ASCII characters in back
1218        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1219        It now works when Unicode Property Support is available.
1220    
1221    5.  In pcretest, an escape such as \x{de} in the data was always generating
1222        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1223        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1224        truncation.
1225    
1226    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1227    
1228    7.  Added two (int) casts to pcregrep when printing the difference of two
1229        pointers, in case they are 64-bit values.
1230    
1231    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1232        test 2 if it fails.
1233    
1234    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1235        and a #define of that name to empty if it is not externally set. This is to
1236        allow users of MSVC to set it if necessary.
1237    
1238    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1239        the convenience functions in the pcre_get.c source file.
1240    
1241    11. An option change at the start of a pattern that had top-level alternatives
1242        could cause overwriting and/or a crash. This command provoked a crash in
1243        some environments:
1244    
1245          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1246    
1247        This potential security problem was recorded as CVE-2008-2371.
1248    
1249    12. For a pattern where the match had to start at the beginning or immediately
1250        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1251        pcre_dfa_exec() could read past the end of the passed subject if there was
1252        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1253        pcretest so that it places the subject at the end of its malloc-ed buffer.
1254    
1255    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1256        exec() might read past the end of the data buffer in UTF-8 mode.
1257    
1258    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1259        the data contained the byte 0x85 as part of a UTF-8 character within its
1260        first line. This applied both to normal and DFA matching.
1261    
1262    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1263        /^[^d]*?$/8 failed to match "abc".
1264    
1265    16. Added a missing copyright notice to pcrecpp_internal.h.
1266    
1267    17. Make it more clear in the documentation that values returned from
1268        pcre_exec() in ovector are byte offsets, not character counts.
1269    
1270    18. Tidied a few places to stop certain compilers from issuing warnings.
1271    
1272    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1273        supplied by Stefan Weber. I made a further small update for 7.8 because
1274        there is a change of source arrangements: the pcre_searchfuncs.c module is
1275        replaced by pcre_ucd.c.
1276    
1277    
1278    Version 7.7 07-May-08
1279  ---------------------  ---------------------
1280    
1281  1.  Applied Craig's patch to sort out a long long problem: "If we can't convert  1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1282      a string to a long long, pretend we don't even have a long long." This is      a string to a long long, pretend we don't even have a long long." This is
1283      done by checking for the strtoq, strtoll, and _strtoi64 functions.      done by checking for the strtoq, strtoll, and _strtoi64 functions.
1284    
1285  2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with  2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1286      pre-7.6 versions, which defined a global no_arg variable instead of putting      pre-7.6 versions, which defined a global no_arg variable instead of putting
1287      it in the RE class.      it in the RE class. (See also #8 below.)
1288    
1289  3.  Remove a line of dead code, identified by coverity and reported by Nuno  3.  Remove a line of dead code, identified by coverity and reported by Nuno
1290      Lopes.      Lopes.
1291    
1292  4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:  4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1293    
1294      (1) The include/exclude patterns were being applied to the whole pathnames      (1) The include/exclude patterns were being applied to the whole pathnames
1295          of files, instead of just to the final components.          of files, instead of just to the final components.
1296    
1297      (2) If there was more than one level of directory, the subdirectories were      (2) If there was more than one level of directory, the subdirectories were
1298          skipped unless they satisfied the include/exclude conditions. This is          skipped unless they satisfied the include/exclude conditions. This is
1299          inconsistent with GNU grep (and could even be seen as contrary to the          inconsistent with GNU grep (and could even be seen as contrary to the
1300          pcregrep specification - which I improved to make it absolutely clear).          pcregrep specification - which I improved to make it absolutely clear).
1301          The action now is always to scan all levels of directory, and just          The action now is always to scan all levels of directory, and just
1302          apply the include/exclude patterns to regular files.          apply the include/exclude patterns to regular files.
1303    
1304  5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used  5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1305      --exclude_dir in the tests to avoid scanning .svn directories.      --exclude_dir in the tests to avoid scanning .svn directories.
1306    
1307  6.  Applied Craig's patch to the QuoteMeta function so that it escapes the  6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1308      NUL character as backslash + 0 rather than backslash + NUL, because PCRE      NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1309      doesn't support NULs in patterns.      doesn't support NULs in patterns.
1310    
1311  7.  Added some missing "const"s to declarations of static tables in  7.  Added some missing "const"s to declarations of static tables in
1312      pcre_compile.c and pcre_dfa_exec.c.      pcre_compile.c and pcre_dfa_exec.c.
1313    
1314    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1315        caused by fix #2  above. (Subsequently also a second patch to fix the
1316        first patch. And a third patch - this was a messy problem.)
1317    
1318    9.  Applied Craig's patch to remove the use of push_back().
1319    
1320    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1321        matching function regexec().
1322    
1323    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1324        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1325        references. PCRE supports relative numbers with this syntax (I don't think
1326        Oniguruma does).
1327    
1328    12. Previously, a group with a zero repeat such as (...){0} was completely
1329        omitted from the compiled regex. However, this means that if the group
1330        was called as a subroutine from elsewhere in the pattern, things went wrong
1331        (an internal error was given). Such groups are now left in the compiled
1332        pattern, with a new opcode that causes them to be skipped at execution
1333        time.
1334    
1335    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1336        to the way PCRE behaves:
1337    
1338        (a) A lone ] character is dis-allowed (Perl treats it as data).
1339    
1340        (b) A back reference to an unmatched subpattern matches an empty string
1341            (Perl fails the current match path).
1342    
1343        (c) A data ] in a character class must be notated as \] because if the
1344            first data character in a class is ], it defines an empty class. (In
1345            Perl it is not possible to have an empty class.) The empty class []
1346            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1347            The negative empty class [^] matches any one character, independently
1348            of the DOTALL setting.
1349    
1350    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1351        non-existent subpattern following a character class starting with ']' and
1352        containing () gave an internal compiling error instead of "reference to
1353        non-existent subpattern". Fortunately, when the pattern did exist, the
1354        compiled code was correct. (When scanning forwards to check for the
1355        existencd of the subpattern, it was treating the data ']' as terminating
1356        the class, so got the count wrong. When actually compiling, the reference
1357        was subsequently set up correctly.)
1358    
1359    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1360        it was being rejected as not supported by pcre_dfa_exec(), even though
1361        other assertions are supported. I have made pcre_dfa_exec() support
1362        (*FAIL).
1363    
1364    16. The implementation of 13c above involved the invention of a new opcode,
1365        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1366        cannot be changed at match time, I realized I could make a small
1367        improvement to matching performance by compiling OP_ALLANY instead of
1368        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1369        on the OP_ANY path.
1370    
1371    17. Compiling pcretest on Windows with readline support failed without the
1372        following two fixes: (1) Make the unistd.h include conditional on
1373        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1374    
1375    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1376        ncurses library to be included for pcretest when ReadLine support is
1377        requested, but also to allow for it to be overridden. This patch came from
1378        Daniel Bergström.
1379    
1380    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1381        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1382        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1383        spotting this.
1384    
1385    
1386  Version 7.6 28-Jan-08  Version 7.6 28-Jan-08

Legend:
Removed from v.327  
changed lines
  Added in v.778

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12