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

Diff of /code/trunk/ChangeLog

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

revision 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 1616 by ph10, Mon Nov 30 17:44:45 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.39 xx-xxxxxx-201x
8    ---------------------------
9    
10    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
11        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
12        misbehaved. This bug was found by the LLVM fuzzer.
13    
14    2.  Similar to the above, if an isolated \E was present between an item and its
15        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
16        bug was found by the LLVM fuzzer.
17    
18    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
19        working correctly in UCP mode.
20    
21    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
22        was set when the pmatch argument was NULL. It now returns REG_INVARG.
23    
24    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
25    
26    6 . An empty \Q\E sequence between an item and its qualifier caused
27        pcre_compile() to misbehave when auto callouts were enabled. This bug was
28        found by the LLVM fuzzer.
29    
30    
31    Version 8.38 23-November-2015
32    -----------------------------
33    
34    1.  If a group that contained a recursive back reference also contained a
35        forward reference subroutine call followed by a non-forward-reference
36        subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
37        compile correct code, leading to undefined behaviour or an internally
38        detected error. This bug was discovered by the LLVM fuzzer.
39    
40    2.  Quantification of certain items (e.g. atomic back references) could cause
41        incorrect code to be compiled when recursive forward references were
42        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
43        This bug was discovered by the LLVM fuzzer.
44    
45    3.  A repeated conditional group whose condition was a reference by name caused
46        a buffer overflow if there was more than one group with the given name.
47        This bug was discovered by the LLVM fuzzer.
48    
49    4.  A recursive back reference by name within a group that had the same name as
50        another group caused a buffer overflow. For example:
51        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
52    
53    5.  A forward reference by name to a group whose number is the same as the
54        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
55        a buffer overflow at compile time. This bug was discovered by the LLVM
56        fuzzer.
57    
58    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
59        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
60    
61    7.  Another buffer overflow bug involved duplicate named groups with a
62        reference between their definition, with a group that reset capture
63        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
64        fixed by always allowing for more memory, even if not needed. (A proper fix
65        is implemented in PCRE2, but it involves more refactoring.)
66    
67    8.  There was no check for integer overflow in subroutine calls such as (?123).
68    
69    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
70        being treated as a literal 'l' instead of causing an error.
71    
72    10. There was a buffer overflow if pcre_exec() was called with an ovector of
73        size 1. This bug was found by american fuzzy lop.
74    
75    11. If a non-capturing group containing a conditional group that could match
76        an empty string was repeated, it was not identified as matching an empty
77        string itself. For example: /^(?:(?(1)x|)+)+$()/.
78    
79    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
80        \a and \e in test subject lines.
81    
82    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
83        instead of the EBCDIC value.
84    
85    14. The handling of \c in an EBCDIC environment has been revised so that it is
86        now compatible with the specification in Perl's perlebcdic page.
87    
88    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
89        ASCII/Unicode. This has now been added to the list of characters that are
90        recognized as white space in EBCDIC.
91    
92    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
93        error (correctly) when used outside a class, but did not give an error
94        within a class.
95    
96    17. \h within a class was incorrectly compiled in EBCDIC environments.
97    
98    18. A pattern with an unmatched closing parenthesis that contained a backward
99        assertion which itself contained a forward reference caused buffer
100        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
101    
102    19. JIT should return with error when the compiled pattern requires more stack
103        space than the maximum.
104    
105    20. A possessively repeated conditional group that could match an empty string,
106        for example, /(?(R))*+/, was incorrectly compiled.
107    
108    21. Fix infinite recursion in the JIT compiler when certain patterns such as
109        /(?:|a|){100}x/ are analysed.
110    
111    22. Some patterns with character classes involving [: and \\ were incorrectly
112        compiled and could cause reading from uninitialized memory or an incorrect
113        error diagnosis.
114    
115    23. Pathological patterns containing many nested occurrences of [: caused
116        pcre_compile() to run for a very long time.
117    
118    24. A conditional group with only one branch has an implicit empty alternative
119        branch and must therefore be treated as potentially matching an empty
120        string.
121    
122    25. If (?R was followed by - or + incorrect behaviour happened instead of a
123        diagnostic.
124    
125    26. Arrange to give up on finding the minimum matching length for overly
126        complex patterns.
127    
128    27. Similar to (4) above: in a pattern with duplicated named groups and an
129        occurrence of (?| it is possible for an apparently non-recursive back
130        reference to become recursive if a later named group with the relevant
131        number is encountered. This could lead to a buffer overflow. Wen Guanxing
132        from Venustech ADLAB discovered this bug.
133    
134    28. If pcregrep was given the -q option with -c or -l, or when handling a
135        binary file, it incorrectly wrote output to stdout.
136    
137    29. The JIT compiler did not restore the control verb head in case of *THEN
138        control verbs. This issue was found by Karl Skomski with a custom LLVM
139        fuzzer.
140    
141    30. Error messages for syntax errors following \g and \k were giving inaccurate
142        offsets in the pattern.
143    
144    31. Added a check for integer overflow in conditions (?(<digits>) and
145        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
146        fuzzer.
147    
148    32. Handling recursive references such as (?2) when the reference is to a group
149        later in the pattern uses code that is very hacked about and error-prone.
150        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
151        give an internal error if it is obvious that compiling has gone wrong.
152    
153    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
154        This issue was found by Karl Skomski with a custom LLVM fuzzer.
155    
156    34. The JIT compiler should restore the control chain for empty possessive
157        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
158    
159    35. Match limit check added to JIT recursion. This issue was found by Karl
160        Skomski with a custom LLVM fuzzer.
161    
162    36. Yet another case similar to 27 above has been circumvented by an
163        unconditional allocation of extra memory. This issue is fixed "properly" in
164        PCRE2 by refactoring the way references are handled. Wen Guanxing
165        from Venustech ADLAB discovered this bug.
166    
167    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
168        with a custom LLVM fuzzer.
169    
170    38. Fixed a corner case of range optimization in JIT.
171    
172    39. An incorrect error "overran compiling workspace" was given if there were
173        exactly enough group forward references such that the last one extended
174        into the workspace safety margin. The next one would have expanded the
175        workspace. The test for overflow was not including the safety margin.
176    
177    40. A match limit issue is fixed in JIT which was found by Karl Skomski
178        with a custom LLVM fuzzer.
179    
180    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
181        work under Windows. (Why has it taken so long for anyone to notice?)
182    
183    42. In a character class such as [\W\p{Any}] where both a negative-type escape
184        ("not a word character") and a property escape were present, the property
185        escape was being ignored.
186    
187    43. Fix crash caused by very long (*MARK) or (*THEN) names.
188    
189    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
190        by a single ASCII character in a class item, was incorrectly compiled in
191        UCP mode. The POSIX class got lost, but only if the single character
192        followed it.
193    
194    45. [:punct:] in UCP mode was matching some characters in the range 128-255
195        that should not have been matched.
196    
197    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
198        class, all characters with code points greater than 255 are in the class.
199        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
200        such as \w are turned into Unicode properties), wide characters were not
201        correctly handled, and could fail to match.
202    
203    
204    Version 8.37 28-April-2015
205    --------------------------
206    
207    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
208        for those parentheses to be closed with whatever has been captured so far.
209        However, it was failing to mark any other groups between the hightest
210        capture so far and the currrent group as "unset". Thus, the ovector for
211        those groups contained whatever was previously there. An example is the
212        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
213    
214    2.  If an assertion condition was quantified with a minimum of zero (an odd
215        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
216    
217    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
218        unrecognized modifier, a crash could occur.
219    
220    4.  An attempt to do global matching in pcretest with a zero-length ovector
221        caused a crash.
222    
223    5.  Fixed a memory leak during matching that could occur for a subpattern
224        subroutine call (recursive or otherwise) if the number of captured groups
225        that had to be saved was greater than ten.
226    
227    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
228        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
229        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
230    
231    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
232        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
233    
234    8. When a pattern is compiled, it remembers the highest back reference so that
235       when matching, if the ovector is too small, extra memory can be obtained to
236       use instead. A conditional subpattern whose condition is a check on a
237       capture having happened, such as, for example in the pattern
238       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
239       setting the highest backreference number. This mattered only if pcre_exec()
240       was called with an ovector that was too small to hold the capture, and there
241       was no other kind of back reference (a situation which is probably quite
242       rare). The effect of the bug was that the condition was always treated as
243       FALSE when the capture could not be consulted, leading to a incorrect
244       behaviour by pcre_exec(). This bug has been fixed.
245    
246    9. A reference to a duplicated named group (either a back reference or a test
247       for being set in a conditional) that occurred in a part of the pattern where
248       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
249       to be incorrectly calculated, leading to overwriting.
250    
251    10. A mutually recursive set of back references such as (\2)(\1) caused a
252        segfault at study time (while trying to find the minimum matching length).
253        The infinite loop is now broken (with the minimum length unset, that is,
254        zero).
255    
256    11. If an assertion that was used as a condition was quantified with a minimum
257        of zero, matching went wrong. In particular, if the whole group had
258        unlimited repetition and could match an empty string, a segfault was
259        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
260        assertions to be quantified, but not if they are being used as conditions,
261        so the above pattern is faulted by Perl. PCRE has now been changed so that
262        it also rejects such patterns.
263    
264    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
265        failed to allow the zero-repeat case if pcre2_exec() was called with an
266        ovector too small to capture the group.
267    
268    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
269        Red Hat Product Security:
270    
271        (a) A crash if /K and /F were both set with the option to save the compiled
272        pattern.
273    
274        (b) Another crash if the option to print captured substrings in a callout
275        was combined with setting a null ovector, for example \O\C+ as a subject
276        string.
277    
278    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
279        forward reference repeated a large (but limited) number of times within a
280        repeated outer group that has a zero minimum quantifier, caused incorrect
281        code to be compiled, leading to the error "internal error:
282        previously-checked referenced subpattern not found" when an incorrect
283        memory address was read. This bug was reported as "heap overflow",
284        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
285        CVE-2015-2325.
286    
287    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
288        call within a group that also contained a recursive back reference caused
289        incorrect code to be compiled. This bug was reported as "heap overflow",
290        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
291        number CVE-2015-2326.
292    
293    24. Computing the size of the JIT read-only data in advance has been a source
294        of various issues, and new ones are still appear unfortunately. To fix
295        existing and future issues, size computation is eliminated from the code,
296        and replaced by on-demand memory allocation.
297    
298    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
299        adjacent to the end of the range, and the range contained characters with
300        more than one other case, caused incorrect behaviour when compiled in UTF
301        mode. In that example, the range a-j was left out of the class.
302    
303    26. Fix JIT compilation of conditional blocks, which assertion
304        is converted to (*FAIL). E.g: /(?(?!))/.
305    
306    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
307        discovered by the LLVM fuzzer.
308    
309    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
310        when this assertion was used as a condition, for example (?(?!)a|b). In
311        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
312        error about an unsupported item.
313    
314    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
315        possessification code could take exponential time to complete. A recursion
316        depth limit of 1000 has been imposed to limit the resources used by this
317        optimization.
318    
319    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
320        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
321        because \S ensures they are all in the class. The code for doing this was
322        interacting badly with the code for computing the amount of space needed to
323        compile the pattern, leading to a buffer overflow. This bug was discovered
324        by the LLVM fuzzer.
325    
326    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
327        other kinds of group caused stack overflow at compile time. This bug was
328        discovered by the LLVM fuzzer.
329    
330    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
331        between a subroutine call and its quantifier was incorrectly compiled,
332        leading to buffer overflow or other errors. This bug was discovered by the
333        LLVM fuzzer.
334    
335    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
336        assertion after (?(. The code was failing to check the character after
337        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
338        was discovered by the LLVM fuzzer.
339    
340    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
341        a fixed maximum following a group that contains a subroutine reference was
342        incorrectly compiled and could trigger buffer overflow. This bug was
343        discovered by the LLVM fuzzer.
344    
345    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
346        caused a stack overflow instead of the diagnosis of a non-fixed length
347        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
348    
349    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
350        (e.g. /(?<=\Ka)/) could make pcregrep loop.
351    
352    37. There was a similar problem to 36 in pcretest for global matches.
353    
354    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
355        and a subsequent item in the pattern caused a non-match, backtracking over
356        the repeated \X did not stop, but carried on past the start of the subject,
357        causing reference to random memory and/or a segfault. There were also some
358        other cases where backtracking after \C could crash. This set of bugs was
359        discovered by the LLVM fuzzer.
360    
361    39. The function for finding the minimum length of a matching string could take
362        a very long time if mutual recursion was present many times in a pattern,
363        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
364        method has been implemented. This infelicity was discovered by the LLVM
365        fuzzer.
366    
367    40. Static linking against the PCRE library using the pkg-config module was
368        failing on missing pthread symbols.
369    
370    
371    Version 8.36 26-September-2014
372    ------------------------------
373    
374    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
375        -Wmissing-field-initializers and -Wunused-parameter.
376    
377    2.  The tests for quantifiers being too big (greater than 65535) were being
378        applied after reading the number, and stupidly assuming that integer
379        overflow would give a negative number. The tests are now applied as the
380        numbers are read.
381    
382    3.  Tidy code in pcre_exec.c where two branches that used to be different are
383        now the same.
384    
385    4.  The JIT compiler did not generate match limit checks for certain
386        bracketed expressions with quantifiers. This may lead to exponential
387        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
388        issue should be resolved now.
389    
390    5.  Fixed an issue, which occures when nested alternatives are optimized
391        with table jumps.
392    
393    6.  Inserted two casts and changed some ints to size_t in the light of some
394        reported 64-bit compiler warnings (Bugzilla 1477).
395    
396    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
397        an empty string, which sometimes were behaving incorrectly in the
398        interpreter (though correctly in the JIT matcher). This pcretest input is
399        an example:
400    
401          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
402          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
403    
404        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
405        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
406        for an empty string was breaking the inner loop and carrying on at a lower
407        level, when possessive repeated groups should always return to a higher
408        level as they have no backtrack points in them. The empty string test now
409        occurs at the outer level.
410    
411    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
412        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
413    
414    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
415        doesn't).
416    
417    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
418        to hold a pattern to be incorrectly computed (too small) when there were
419        named back references to duplicated names. This could cause "internal
420        error: code overflow" or "double free or corruption" or other memory
421        handling errors.
422    
423    11. When named subpatterns had the same prefixes, back references could be
424        confused. For example, in this pattern:
425    
426          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
427    
428        the reference to 'Name' was incorrectly treated as a reference to a
429        duplicate name.
430    
431    12. A pattern such as /^s?c/mi8 where the optional character has more than
432        one "other case" was incorrectly compiled such that it would only try to
433        match starting at "c".
434    
435    13. When a pattern starting with \s was studied, VT was not included in the
436        list of possible starting characters; this should have been part of the
437        8.34/18 patch.
438    
439    14. If a character class started [\Qx]... where x is any character, the class
440        was incorrectly terminated at the ].
441    
442    15. If a pattern that started with a caseless match for a character with more
443        than one "other case" was studied, PCRE did not set up the starting code
444        unit bit map for the list of possible characters. Now it does. This is an
445        optimization improvement, not a bug fix.
446    
447    16. The Unicode data tables have been updated to Unicode 7.0.0.
448    
449    17. Fixed a number of memory leaks in pcregrep.
450    
451    18. Avoid a compiler warning (from some compilers) for a function call with
452        a cast that removes "const" from an lvalue by using an intermediate
453        variable (to which the compiler does not object).
454    
455    19. Incorrect code was compiled if a group that contained an internal recursive
456        back reference was optional (had quantifier with a minimum of zero). This
457        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
458        caused segmentation faults because of stack overflows at compile time.
459    
460    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
461        group that is quantified with an indefinite repeat, caused a compile-time
462        loop which used up all the system stack and provoked a segmentation fault.
463        This was not the same bug as 19 above.
464    
465    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
466        Patch by Mike Frysinger.
467    
468    
469    Version 8.35 04-April-2014
470    --------------------------
471    
472    1.  A new flag is set, when property checks are present in an XCLASS.
473        When this flag is not set, PCRE can perform certain optimizations
474        such as studying these XCLASS-es.
475    
476    2.  The auto-possessification of character sets were improved: a normal
477        and an extended character set can be compared now. Furthermore
478        the JIT compiler optimizes more character set checks.
479    
480    3.  Got rid of some compiler warnings for potentially uninitialized variables
481        that show up only when compiled with -O2.
482    
483    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
484        of a match later then the end of the match. The pcretest program was not
485        handling the case sensibly - it was outputting from the start to the next
486        binary zero. It now reports this situation in a message, and outputs the
487        text from the end to the start.
488    
489    5.  Fast forward search is improved in JIT. Instead of the first three
490        characters, any three characters with fixed position can be searched.
491        Search order: first, last, middle.
492    
493    6.  Improve character range checks in JIT. Characters are read by an inprecise
494        function now, which returns with an unknown value if the character code is
495        above a certain threshold (e.g: 256). The only limitation is that the value
496        must be bigger than the threshold as well. This function is useful when
497        the characters above the threshold are handled in the same way.
498    
499    7.  The macros whose names start with RAWUCHAR are placeholders for a future
500        mode in which only the bottom 21 bits of 32-bit data items are used. To
501        make this more memorable for those maintaining the code, the names have
502        been changed to start with UCHAR21, and an extensive comment has been added
503        to their definition.
504    
505    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
506        to the export list in Makefile.am (they were accidentally omitted from the
507        8.34 tarball).
508    
509    9.  The informational output from pcretest used the phrase "starting byte set"
510        which is inappropriate for the 16-bit and 32-bit libraries. As the output
511        for "first char" and "need char" really means "non-UTF-char", I've changed
512        "byte" to "char", and slightly reworded the output. The documentation about
513        these values has also been (I hope) clarified.
514    
515    10. Another JIT related optimization: use table jumps for selecting the correct
516        backtracking path, when more than four alternatives are present inside a
517        bracket.
518    
519    11. Empty match is not possible, when the minimum length is greater than zero,
520        and there is no \K in the pattern. JIT should avoid empty match checks in
521        such cases.
522    
523    12. In a caseless character class with UCP support, when a character with more
524        than one alternative case was not the first character of a range, not all
525        the alternative cases were added to the class. For example, s and \x{17f}
526        are both alternative cases for S: the class [RST] was handled correctly,
527        but [R-T] was not.
528    
529    13. The configure.ac file always checked for pthread support when JIT was
530        enabled. This is not used in Windows, so I have put this test inside a
531        check for the presence of windows.h (which was already tested for).
532    
533    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
534        The algorithm provides a way to skip certain starting offsets, and usually
535        faster than linear prefix searches.
536    
537    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
538        as for 'fr_FR' and 'french'. For some reason, however, it then used the
539        Windows-specific input and output files, which have 'french' screwed in.
540        So this could never have worked. One of the problems with locales is that
541        they aren't always the same. I have now updated RunTest so that it checks
542        the output of the locale test (test 3) against three different output
543        files, and it allows the test to pass if any one of them matches. With luck
544        this should make the test pass on some versions of Solaris where it was
545        failing. Because of the uncertainty, the script did not used to stop if
546        test 3 failed; it now does. If further versions of a French locale ever
547        come to light, they can now easily be added.
548    
549    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
550        there was a message during ./configure, but it did not stop. This now
551        provokes an error. The invalid example in README has been corrected.
552        If a value less than the minimum is given, the minimum value has always
553        been used, but now a warning is given.
554    
555    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
556        was a bug in the test system, which is now fixed. Also, the list of various
557        configurations that are tested for each release did not have one with both
558        16/32 bits and --enable-bar-anycrlf. It now does.
559    
560    18. pcretest was missing "-C bsr" for displaying the \R default setting.
561    
562    19. Little endian PowerPC systems are supported now by the JIT compiler.
563    
564    20. The fast forward newline mechanism could enter to an infinite loop on
565        certain invalid UTF-8 input. Although we don't support these cases
566        this issue can be fixed by a performance optimization.
567    
568    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
569        not take account if existing stack usage. There is now a new global
570        variable called pcre_stack_guard that can be set to point to an external
571        function to check stack availability. It is called at the start of
572        processing every parenthesized group.
573    
574    22. A typo in the code meant that in ungreedy mode the max/min qualifier
575        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
576        not match "ab".
577    
578    23. When UTF was disabled, the JIT program reported some incorrect compile
579        errors. These messages are silenced now.
580    
581    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
582        compiler.
583    
584    25. Change all the temporary files used in RunGrepTest to be different to those
585        used by RunTest so that the tests can be run simultaneously, for example by
586        "make -j check".
587    
588    
589    Version 8.34 15-December-2013
590    -----------------------------
591    
592    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
593        executable memory. Patch inspired by Carsten Klein.
594    
595    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
596        this macro is never tested and has no effect, because the work to support
597        coverage involves only compiling and linking options and special targets in
598        the Makefile. The comment in config.h implied that defining the macro would
599        enable coverage support, which is totally false. There was also support for
600        setting this macro in the CMake files (my fault, I just copied it from
601        configure). SUPPORT_GCOV has now been removed.
602    
603    3.  Make a small performance improvement in strlen16() and strlen32() in
604        pcretest.
605    
606    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
607        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
608        these cases). There was also one in pcretest.c.
609    
610    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
611    
612    6.  In UTF mode, the code for checking whether a group could match an empty
613        string (which is used for indefinitely repeated groups to allow for
614        breaking an infinite loop) was broken when the group contained a repeated
615        negated single-character class with a character that occupied more than one
616        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
617        UTF-8 mode). The effect was undefined: the group might or might not be
618        deemed as matching an empty string, or the program might have crashed.
619    
620    7.  The code for checking whether a group could match an empty string was not
621        recognizing that \h, \H, \v, \V, and \R must match a character.
622    
623    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
624        an empty string. If it can, pcretest shows this in its information output.
625    
626    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
627        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
628        matched by pcre_exec() without using JIT:
629    
630        (a) If the rest of the pattern did not match after a maximal run of
631            grapheme clusters, the code for backing up to try with fewer of them
632            did not always back up over a full grapheme when characters that do not
633            have the modifier quality were involved, e.g. Hangul syllables.
634    
635        (b) If the match point in a subject started with modifier character, and
636            there was no match, the code could incorrectly back up beyond the match
637            point, and potentially beyond the first character in the subject,
638            leading to a segfault or an incorrect match result.
639    
640    10. A conditional group with an assertion condition could lead to PCRE
641        recording an incorrect first data item for a match if no other first data
642        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
643        first data item, and therefore matched "ca" after "c" instead of at the
644        start.
645    
646    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
647        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
648    
649    12. The source of pcregrep now includes z/OS-specific code so that it can be
650        compiled for z/OS as part of the special z/OS distribution.
651    
652    13. Added the -T and -TM options to pcretest.
653    
654    14. The code in pcre_compile.c for creating the table of named capturing groups
655        has been refactored. Instead of creating the table dynamically during the
656        actual compiling pass, the information is remembered during the pre-compile
657        pass (on the stack unless there are more than 20 named groups, in which
658        case malloc() is used) and the whole table is created before the actual
659        compile happens. This has simplified the code (it is now nearly 150 lines
660        shorter) and prepared the way for better handling of references to groups
661        with duplicate names.
662    
663    15. A back reference to a named subpattern when there is more than one of the
664        same name now checks them in the order in which they appear in the pattern.
665        The first one that is set is used for the reference. Previously only the
666        first one was inspected. This change makes PCRE more compatible with Perl.
667    
668    16. Unicode character properties were updated from Unicode 6.3.0.
669    
670    17. The compile-time code for auto-possessification has been refactored, based
671        on a patch by Zoltan Herczeg. It now happens after instead of during
672        compilation. The code is cleaner, and more cases are handled. The option
673        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
674        options in pcretest are provided to set it. It can also be set by
675        (*NO_AUTO_POSSESS) at the start of a pattern.
676    
677    18. The character VT has been added to the default ("C" locale) set of
678        characters that match \s and are generally treated as white space,
679        following this same change in Perl 5.18. There is now no difference between
680        "Perl space" and "POSIX space". Whether VT is treated as white space in
681        other locales depends on the locale.
682    
683    19. The code for checking named groups as conditions, either for being set or
684        for being recursed, has been refactored (this is related to 14 and 15
685        above). Processing unduplicated named groups should now be as fast at
686        numerical groups, and processing duplicated groups should be faster than
687        before.
688    
689    20. Two patches to the CMake build system, by Alexander Barkov:
690    
691          (1) Replace the "source" command by "." in CMakeLists.txt because
692              "source" is a bash-ism.
693    
694          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
695              without these the CMake build does not work on Solaris.
696    
697    21. Perl has changed its handling of \8 and \9. If there is no previously
698        encountered capturing group of those numbers, they are treated as the
699        literal characters 8 and 9 instead of a binary zero followed by the
700        literals. PCRE now does the same.
701    
702    22. Following Perl, added \o{} to specify codepoints in octal, making it
703        possible to specify values greater than 0777 and also making them
704        unambiguous.
705    
706    23. Perl now gives an error for missing closing braces after \x{... instead of
707        treating the string as literal. PCRE now does the same.
708    
709    24. RunTest used to grumble if an inappropriate test was selected explicitly,
710        but just skip it when running all tests. This make it awkward to run ranges
711        of tests when one of them was inappropriate. Now it just skips any
712        inappropriate tests, as it always did when running all tests.
713    
714    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
715        character types such as \d or \w, too many callouts were inserted, and the
716        data that they returned was rubbish.
717    
718    26. In UCP mode, \s was not matching two of the characters that Perl matches,
719        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
720        were matched by \h. The code has now been refactored so that the lists of
721        the horizontal and vertical whitespace characters used for \h and \v (which
722        are defined only in one place) are now also used for \s.
723    
724    27. Add JIT support for the 64 bit TileGX architecture.
725        Patch by Jiong Wang (Tilera Corporation).
726    
727    28. Possessive quantifiers for classes (both explicit and automatically
728        generated) now use special opcodes instead of wrapping in ONCE brackets.
729    
730    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
731        (because it's meaningless), this was not happening when PCRE_CASELESS was
732        set. Not wrong, but inefficient.
733    
734    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
735        properties for \w, \d, etc) is present in a test regex. Otherwise if the
736        test contains no characters greater than 255, Perl doesn't realise it
737        should be using Unicode semantics.
738    
739    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
740        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
741        does in Unicode mode.
742    
743    32. Added the "forbid" facility to pcretest so that putting tests into the
744        wrong test files can sometimes be quickly detected.
745    
746    33. There is now a limit (default 250) on the depth of nesting of parentheses.
747        This limit is imposed to control the amount of system stack used at compile
748        time. It can be changed at build time by --with-parens-nest-limit=xxx or
749        the equivalent in CMake.
750    
751    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
752        errors. Perl warns for these when in warning mode, but PCRE has no facility
753        for giving warnings.
754    
755    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
756        However, this was not working for (?!) because it is optimized to (*FAIL),
757        for which PCRE does not allow quantifiers. The optimization is now disabled
758        when a quantifier follows (?!). I can't see any use for this, but it makes
759        things uniform.
760    
761    36. Perl no longer allows group names to start with digits, so I have made this
762        change also in PCRE. It simplifies the code a bit.
763    
764    37. In extended mode, Perl ignores spaces before a + that indicates a
765        possessive quantifier. PCRE allowed a space before the quantifier, but not
766        before the possessive +. It now does.
767    
768    38. The use of \K (reset reported match start) within a repeated possessive
769        group such as (a\Kb)*+ was not working.
770    
771    40. Document that the same character tables must be used at compile time and
772        run time, and that the facility to pass tables to pcre_exec() and
773        pcre_dfa_exec() is for use only with saved/restored patterns.
774    
775    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
776        features for Builds with MSVC:
777    
778        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
779           builds). The .rc files can be used to set FileDescription and many other
780           attributes.
781    
782        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
783           This allows higher-level build scripts which want .pdb files to avoid
784           hard-coding the exact files needed."
785    
786    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
787        mean "start of word" and "end of word", respectively, as a transition aid.
788    
789    43. A minimizing repeat of a class containing codepoints greater than 255 in
790        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
791        compiled to use the heap for recursion.
792    
793    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
794        is configured.
795    
796    
797    Version 8.33 28-May-2013
798    ------------------------
799    
800    1.  Added 'U' to some constants that are compared to unsigned integers, to
801        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
802        variables that are added to signed variables, to ensure the result is
803        signed and can be negated.
804    
805    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
806        CMake config files.
807    
808    3.  Revise the creation of config.h.generic so that all boolean macros are
809        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
810        overriding via -D on the command line possible.
811    
812    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
813        to unsigned int is reported to make a quite noticeable speed difference in
814        a specific Windows environment. Testing on Linux did also appear to show
815        some benefit (and it is clearly not harmful). Also fixed the definition of
816        Xop which should be unsigned.
817    
818    5.  Related to (4), changing the definition of the intermediate variable cc
819        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
820        improvements.
821    
822    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
823        unnecessary spaces.
824    
825    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
826        and later.
827    
828    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
829    
830    9.  Optimizing fast_forward_start_bits in JIT.
831    
832    10. Adding support for callouts in JIT, and fixing some issues revealed
833        during this work. Namely:
834    
835        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
836    
837        (b) Minimum length was not checked before the matching is started.
838    
839    11. The value of capture_last that is passed to callouts was incorrect in some
840        cases when there was a capture on one path that was subsequently abandoned
841        after a backtrack. Also, the capture_last value is now reset after a
842        recursion, since all captures are also reset in this case.
843    
844    12. The interpreter no longer returns the "too many substrings" error in the
845        case when an overflowing capture is in a branch that is subsequently
846        abandoned after a backtrack.
847    
848    13. In the pathological case when an offset vector of size 2 is used, pcretest
849        now prints out the matched string after a yield of 0 or 1.
850    
851    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
852        Only supported by the JIT compiler at the moment.
853    
854    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
855    
856    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
857        offset of the starting point of the matching process, provided the offsets
858        vector is large enough.
859    
860    17. The \A escape now records a lookbehind value of 1, though its execution
861        does not actually inspect the previous character. This is to ensure that,
862        in partial multi-segment matching, at least one character from the old
863        segment is retained when a new segment is processed. Otherwise, if there
864        are no lookbehinds in the pattern, \A might match incorrectly at the start
865        of a new segment.
866    
867    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
868    
869    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
870        gives some modest performance improvement in 8-bit mode.
871    
872    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
873        be expressed in certain programming languages using Universal Character
874        Names.
875    
876    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
877        which points out that "non characters" are not "characters that may not
878        appear in Unicode strings" but rather "characters that are reserved for
879        internal use and have only local meaning".
880    
881    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
882        there was a conditional group that depended on an assertion, if the
883        assertion was false, the callout that immediately followed the alternation
884        in the condition was skipped when pcre_exec() was used for matching.
885    
886    23. Allow an explicit callout to be inserted before an assertion that is the
887        condition for a conditional group, for compatibility with automatic
888        callouts, which always insert a callout at this point.
889    
890    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
891        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
892    
893    25. (*PRUNE) is now supported by the JIT compiler.
894    
895    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
896    
897    27. Fix the case where there are two or more SKIPs with arguments that may be
898        ignored.
899    
900    28. (*SKIP) is now supported by the JIT compiler.
901    
902    29. (*THEN) is now supported by the JIT compiler.
903    
904    30. Update RunTest with additional test selector options.
905    
906    31. The way PCRE handles backtracking verbs has been changed in two ways.
907    
908        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
909        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
910        backtracking. In some cases this makes it more Perl-compatible, but Perl's
911        rather obscure rules do not always do the same thing.
912    
913        (2) Previously, backtracking verbs were confined within assertions. This is
914        no longer the case for positive assertions, except for (*ACCEPT). Again,
915        this sometimes improves Perl compatibility, and sometimes does not.
916    
917    32. A number of tests that were in test 2 because Perl did things differently
918        have been moved to test 1, because either Perl or PCRE has changed, and
919        these tests are now compatible.
920    
921    32. Backtracking control verbs are now handled in the same way in JIT and
922        interpreter.
923    
924    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
925        contained a forward subroutine reference caused a compile error.
926    
927    34. Auto-detect and optimize limited repetitions in JIT.
928    
929    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
930        blocking (*UTF) etc.
931    
932    36. In the interpreter, maximizing pattern repetitions for characters and
933        character types now use tail recursion, which reduces stack usage.
934    
935    37. The value of the max lookbehind was not correctly preserved if a compiled
936        and saved regex was reloaded on a host of different endianness.
937    
938    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
939        of the compiled pattern block, expand the flags field from 16 to 32 bits
940        because it was almost full.
941    
942    39. Try madvise first before posix_madvise.
943    
944    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
945        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
946        original change locked out all matches of empty strings. This has been
947        changed so that one match of an empty string per line is recognized.
948        Subsequent searches on the same line (for colouring or for --only-matching,
949        for example) do not recognize empty strings.
950    
951    41. Applied a user patch to fix a number of spelling mistakes in comments.
952    
953    42. Data lines longer than 65536 caused pcretest to crash.
954    
955    43. Clarified the data type for length and startoffset arguments for pcre_exec
956        and pcre_dfa_exec in the function-specific man pages, where they were
957        explicitly stated to be in bytes, never having been updated. I also added
958        some clarification to the pcreapi man page.
959    
960    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
961        a segmentation fault.
962    
963    
964    Version 8.32 30-November-2012
965    -----------------------------
966    
967    1.  Improved JIT compiler optimizations for first character search and single
968        character iterators.
969    
970    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
971        Patch by Daniel Richard G.
972    
973    3.  Single character iterator optimizations in the JIT compiler.
974    
975    4.  Improved JIT compiler optimizations for character ranges.
976    
977    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
978        Reported by Giuseppe D'Angelo.
979    
980    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
981        of a line, was being set incorrectly in cases where .* appeared inside
982        atomic brackets at the start of a pattern, or where there was a subsequent
983        *PRUNE or *SKIP.
984    
985    7.  Improved instruction cache flush for POWER/PowerPC.
986        Patch by Daniel Richard G.
987    
988    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
989        grep:
990    
991        (a) There is now no limit to the number of patterns to be matched.
992    
993        (b) An error is given if a pattern is too long.
994    
995        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
996            are now supported.
997    
998        (d) --exclude-from and --include-from (multiple use) have been added.
999    
1000        (e) Exclusions and inclusions now apply to all files and directories, not
1001            just to those obtained from scanning a directory recursively.
1002    
1003        (f) Multiple uses of -f and --file-list are now supported.
1004    
1005        (g) In a Windows environment, the default for -d has been changed from
1006            "read" (the GNU grep default) to "skip", because otherwise the presence
1007            of a directory in the file list provokes an error.
1008    
1009        (h) The documentation has been revised and clarified in places.
1010    
1011    9.  Improve the matching speed of capturing brackets.
1012    
1013    10. Changed the meaning of \X so that it now matches a Unicode extended
1014        grapheme cluster.
1015    
1016    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1017        out POSIX threads when JIT support is configured.
1018    
1019    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1020    
1021    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1022        structure could go wrong in environments where size_t is not the same size
1023        as int.
1024    
1025    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1026        set.
1027    
1028    15. The EBCDIC support had decayed; later updates to the code had included
1029        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1030        general tidy up of EBCDIC-related issues, and the documentation was also
1031        not quite right. There is now a test that can be run on ASCII systems to
1032        check some of the EBCDIC-related things (but is it not a full test).
1033    
1034    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1035        in a small tidy to the code.
1036    
1037    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1038    
1039    18. If the --only-matching (-o) option in pcregrep is specified multiple
1040        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1041        substrings matched by the 1st and 2nd capturing parentheses. A separating
1042        string can be specified by --om-separator (default empty).
1043    
1044    19. Improving the first n character searches.
1045    
1046    20. Turn case lists for horizontal and vertical white space into macros so that
1047        they are defined only once.
1048    
1049    21. This set of changes together give more compatible Unicode case-folding
1050        behaviour for characters that have more than one other case when UCP
1051        support is available.
1052    
1053        (a) The Unicode property table now has offsets into a new table of sets of
1054            three or more characters that are case-equivalent. The MultiStage2.py
1055            script that generates these tables (the pcre_ucd.c file) now scans
1056            CaseFolding.txt instead of UnicodeData.txt for character case
1057            information.
1058    
1059        (b) The code for adding characters or ranges of characters to a character
1060            class has been abstracted into a generalized function that also handles
1061            case-independence. In UTF-mode with UCP support, this uses the new data
1062            to handle characters with more than one other case.
1063    
1064        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1065            whose other case is greater than 256 are now correctly matched
1066            caselessly. Previously, the high codepoint matched the low one, but not
1067            vice versa.
1068    
1069        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1070            of the new class addition function, using character lists defined as
1071            macros alongside the case definitions of 20 above.
1072    
1073        (e) Caseless back references now work with characters that have more than
1074            one other case.
1075    
1076        (f) General caseless matching of characters with more than one other case
1077            is supported.
1078    
1079    22. Unicode character properties were updated from Unicode 6.2.0
1080    
1081    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1082    
1083    24. Add support for 32-bit character strings, and UTF-32
1084    
1085    25. Major JIT compiler update (code refactoring and bugfixing).
1086        Experimental Sparc 32 support is added.
1087    
1088    26. Applied a modified version of Daniel Richard G's patch to create
1089        pcre.h.generic and config.h.generic by "make" instead of in the
1090        PrepareRelease script.
1091    
1092    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1093        pcre_compile.c when checking for a zero character.
1094    
1095    28. Introducing a native interface for JIT. Through this interface, the compiled
1096        machine code can be directly executed. The purpose of this interface is to
1097        provide fast pattern matching, so several sanity checks are not performed.
1098        However, feature tests are still performed. The new interface provides
1099        1.4x speedup compared to the old one.
1100    
1101    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1102        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1103        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1104    
1105    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1106        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1107        the "old" RFC 2279). Instead, it ended up passing a negative length to
1108        pcre_exec().
1109    
1110    31. Add support for GCC's visibility feature to hide internal functions.
1111    
1112    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1113        "unknown -C option" after outputting 0 or 1.
1114    
1115    33. There is now support for generating a code coverage report for the test
1116        suite in environments where gcc is the compiler and lcov is installed. This
1117        is mainly for the benefit of the developers.
1118    
1119    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1120        unaddressable using valgrind annotations, allowing valgrind to detect
1121        invalid memory accesses. This is mainly for the benefit of the developers.
1122    
1123    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1124    
1125    26. Give configure error if --enable-cpp but no C++ compiler found.
1126    
1127    
1128    Version 8.31 06-July-2012
1129    -------------------------
1130    
1131    1.  Fixing a wrong JIT test case and some compiler warnings.
1132    
1133    2.  Removed a bashism from the RunTest script.
1134    
1135    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1136        to unsigned type, result still unsigned" that was given by an MS compiler
1137        on encountering the code "-sizeof(xxx)".
1138    
1139    4.  Partial matching support is added to the JIT compiler.
1140    
1141    5.  Fixed several bugs concerned with partial matching of items that consist
1142        of more than one character:
1143    
1144        (a) /^(..)\1/ did not partially match "aba" because checking references was
1145            done on an "all or nothing" basis. This also applied to repeated
1146            references.
1147    
1148        (b) \R did not give a hard partial match if \r was found at the end of the
1149            subject.
1150    
1151        (c) \X did not give a hard partial match after matching one or more
1152            characters at the end of the subject.
1153    
1154        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1155            a partial match for the string "\r".
1156    
1157        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1158            a partial match for a CR character at the end of the subject string.
1159    
1160    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1161        running pcretest, the text "(JIT)" added to the output whenever JIT is
1162        actually used to run the match.
1163    
1164    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1165        or /S+[+] with a digit between 1 and 7.
1166    
1167    8.  OP_NOT now supports any UTF character not just single-byte ones.
1168    
1169    9.  (*MARK) control verb is now supported by the JIT compiler.
1170    
1171    10. The command "./RunTest list" lists the available tests without actually
1172        running any of them. (Because I keep forgetting what they all are.)
1173    
1174    11. Add PCRE_INFO_MAXLOOKBEHIND.
1175    
1176    12. Applied a (slightly modified) user-supplied patch that improves performance
1177        when the heap is used for recursion (compiled with --disable-stack-for-
1178        recursion). Instead of malloc and free for each heap frame each time a
1179        logical recursion happens, frames are retained on a chain and re-used where
1180        possible. This sometimes gives as much as 30% improvement.
1181    
1182    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1183        call.
1184    
1185    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1186    
1187    15. It is now possible to link pcretest with libedit as an alternative to
1188        libreadline.
1189    
1190    16. (*COMMIT) control verb is now supported by the JIT compiler.
1191    
1192    17. The Unicode data tables have been updated to Unicode 6.1.0.
1193    
1194    18. Added --file-list option to pcregrep.
1195    
1196    19. Added binary file support to pcregrep, including the -a, --binary-files,
1197        -I, and --text options.
1198    
1199    20. The madvise function is renamed for posix_madvise for QNX compatibility
1200        reasons. Fixed by Giuseppe D'Angelo.
1201    
1202    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1203    
1204    22. Changed the option for creating symbolic links for 16-bit man pages from
1205        -s to -sf so that re-installing does not cause issues.
1206    
1207    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1208    
1209    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1210        matches in certain environments (the workspace was not being correctly
1211        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1212        some of the workspace data at the beginning of a restart.
1213    
1214    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1215        was not doing so when it should - probably a typo introduced by SVN 528
1216        (change 8.10/14).
1217    
1218    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1219        \w+ when the character tables indicated that \x{c4} was a word character.
1220        There were several related cases, all because the tests for doing a table
1221        lookup were testing for characters less than 127 instead of 255.
1222    
1223    27. If a pattern contains capturing parentheses that are not used in a match,
1224        their slots in the ovector are set to -1. For those that are higher than
1225        any matched groups, this happens at the end of processing. In the case when
1226        there were back references that the ovector was too small to contain
1227        (causing temporary malloc'd memory to be used during matching), and the
1228        highest capturing number was not used, memory off the end of the ovector
1229        was incorrectly being set to -1. (It was using the size of the temporary
1230        memory instead of the true size.)
1231    
1232    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1233        ovector size, it uses memory at the end of the block that it has got.
1234    
1235    29. Check for an overlong MARK name and give an error at compile time. The
1236        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1237    
1238    30. JIT compiler update.
1239    
1240    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1241        Rill for the patch.
1242    
1243    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1244    
1245    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1246    
1247    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1248        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1249        was enabled.
1250    
1251    35. Improve JIT code generation for greedy plus quantifier.
1252    
1253    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1254        1 to "aa" instead of to an empty string. The bug affected repeated groups
1255        that could potentially match an empty string.
1256    
1257    37. Optimizing single character iterators in JIT.
1258    
1259    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1260        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1261        codepoints that are too big for the mode are faulted, and in a UTF mode,
1262        disallowed codepoints are also faulted.
1263    
1264    39. If PCRE was compiled with UTF support, in three places in the DFA
1265        matcher there was code that should only have been obeyed in UTF mode, but
1266        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1267        processing when bytes with values greater than 127 were present. In 16-bit
1268        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1269        both cases the values are those that cannot be the first data item in a UTF
1270        character. The three items that might have provoked this were recursions,
1271        possessively repeated groups, and atomic groups.
1272    
1273    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1274        and pcregrep, because some OS require shared objects to be explicitly
1275        passed to ld, causing the link step to fail if they are not.
1276    
1277    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1278        mode, patterns that started with \h* or \R* might be incorrectly matched.
1279    
1280    
1281    Version 8.30 04-February-2012
1282    -----------------------------
1283    
1284    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1285        name is defined in ctype.h.
1286    
1287    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1288        only in quite long subpatterns.
1289    
1290    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1291        since it was replaced by pcre_fullinfo() in February 2000.
1292    
1293    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1294        match a (*MARK), and the match failed at the start of the subject, a
1295        reference to memory before the start of the subject could occur. This bug
1296        was introduced by fix 17 of release 8.21.
1297    
1298    5.  A reference to an unset group with zero minimum repetition was giving
1299        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1300        /(another)?(\1?)test/ matched against "hello world test". This bug was
1301        introduced in release 8.13.
1302    
1303    6.  Add support for 16-bit character strings (a large amount of work involving
1304        many changes and refactorings).
1305    
1306    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1307        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1308        from a file.
1309    
1310    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1311        rounding is not applied in this particular case).
1312    
1313    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1314        if they appear, or are escaped, in patterns.
1315    
1316    10. Get rid of a number of -Wunused-but-set-variable warnings.
1317    
1318    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1319        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1320        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1321        also returns the mark "x". This bug applied to capturing parentheses,
1322        non-capturing parentheses, and atomic parentheses. It also applied to some
1323        assertions.
1324    
1325    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1326        information out of configure.ac instead of relying on pcre.h.generic, which
1327        is not stored in the repository.
1328    
1329    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1330        -lreadline.
1331    
1332    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1333    
1334    15. Applied Graycode's patch to put the top-level frame on the stack rather
1335        than the heap when not using the stack for recursion. This gives a
1336        performance improvement in many cases when recursion is not deep.
1337    
1338    16. Experimental code added to "pcretest -C" to output the stack frame size.
1339    
1340    
1341    Version 8.21 12-Dec-2011
1342    ------------------------
1343    
1344    1.  Updating the JIT compiler.
1345    
1346    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1347        are added as well.
1348    
1349    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1350        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1351        calling _pcre_jit_exec. Some extra comments are added.
1352    
1353    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1354        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1355        was introduced by change 18 for 8.20.
1356    
1357    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1358        ECMA-262 standard.
1359    
1360    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1361        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1362        This bug was probably introduced by change 9 of 8.13.
1363    
1364    7.  While fixing 6 above, I noticed that a number of other items were being
1365        incorrectly rejected as "not fixed length". This arose partly because newer
1366        opcodes had not been added to the fixed-length checking code. I have (a)
1367        corrected the bug and added tests for these items, and (b) arranged for an
1368        error to occur if an unknown opcode is encountered while checking for fixed
1369        length instead of just assuming "not fixed length". The items that were
1370        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1371        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1372        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1373    
1374    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1375        being incorrectly compiled and would have given unpredicatble results.
1376    
1377    9.  A possessively repeated subpattern with minimum repeat count greater than
1378        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1379        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1380        the first (A) could occur when it should not.
1381    
1382    10. Add a cast and remove a redundant test from the code.
1383    
1384    11. JIT should use pcre_malloc/pcre_free for allocation.
1385    
1386    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1387        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1388        is anything other than /usr, -L is still shown).
1389    
1390    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1391    
1392    14. Perl does not support \N without a following name in a [] class; PCRE now
1393        also gives an error.
1394    
1395    15. If a forward reference was repeated with an upper limit of around 2000,
1396        it caused the error "internal error: overran compiling workspace". The
1397        maximum number of forward references (including repeats) was limited by the
1398        internal workspace, and dependent on the LINK_SIZE. The code has been
1399        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1400        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1401        of around 200,000 forward references. While doing this, I also speeded up
1402        the filling in of repeated forward references.
1403    
1404    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1405        incorrectly expecting the subject to contain another "a" after the start.
1406    
1407    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1408        in the match, the SKIP should be ignored. This was not happening; instead
1409        the SKIP was being treated as NOMATCH. For patterns such as
1410        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1411        tested.
1412    
1413    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1414        now much more compatible with Perl, in particular in cases where the result
1415        is a non-match for a non-anchored pattern. For example, if
1416        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1417        "m", where previously it did not return a name. A side effect of this
1418        change is that for partial matches, the last encountered mark name is
1419        returned, as for non matches. A number of tests that were previously not
1420        Perl-compatible have been moved into the Perl-compatible test files. The
1421        refactoring has had the pleasing side effect of removing one argument from
1422        the match() function, thus reducing its stack requirements.
1423    
1424    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1425        subsequent uses of /S (without +) incorrectly behaved like /S+.
1426    
1427    21. Retrieve executable code size support for the JIT compiler and fixing
1428        some warnings.
1429    
1430    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1431        not work when the shorter character appeared right at the end of the
1432        subject string.
1433    
1434    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1435        systems.
1436    
1437    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1438        output it when the /M option is used in pcretest.
1439    
1440    25. The CheckMan script was not being included in the distribution. Also, added
1441        an explicit "perl" to run Perl scripts from the PrepareRelease script
1442        because this is reportedly needed in Windows.
1443    
1444    26. If study data was being save in a file and studying had not found a set of
1445        "starts with" bytes for the pattern, the data written to the file (though
1446        never used) was taken from uninitialized memory and so caused valgrind to
1447        complain.
1448    
1449    27. Updated RunTest.bat as provided by Sheri Pierce.
1450    
1451    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1452    
1453    29. Computation of memory usage for the table of capturing group names was
1454        giving an unnecessarily large value.
1455    
1456    
1457    Version 8.20 21-Oct-2011
1458    ------------------------
1459    
1460    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1461        a POSIX class. After further experiments with Perl, which convinced me that
1462        Perl has bugs and confusions, a closing square bracket is no longer allowed
1463        in a POSIX name. This bug also affected patterns with classes that started
1464        with full stops.
1465    
1466    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1467        captured substring, but while checking the failing first alternative,
1468        substring 1 is temporarily captured. If the output vector supplied to
1469        pcre_exec() was not big enough for this capture, the yield of the function
1470        was still zero ("insufficient space for captured substrings"). This cannot
1471        be totally fixed without adding another stack variable, which seems a lot
1472        of expense for a edge case. However, I have improved the situation in cases
1473        such as /(a)(b)x|abc/ matched against "abc", where the return code
1474        indicates that fewer than the maximum number of slots in the ovector have
1475        been set.
1476    
1477    3.  Related to (2) above: when there are more back references in a pattern than
1478        slots in the output vector, pcre_exec() uses temporary memory during
1479        matching, and copies in the captures as far as possible afterwards. It was
1480        using the entire output vector, but this conflicts with the specification
1481        that only 2/3 is used for passing back captured substrings. Now it uses
1482        only the first 2/3, for compatibility. This is, of course, another edge
1483        case.
1484    
1485    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1486        main code base, and can be used by building with --enable-jit. When this is
1487        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1488        runtime --no-jit option is given.
1489    
1490    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1491        ovector, the return from the function was zero, implying that there were
1492        other matches that did not fit. The correct "exactly full" value is now
1493        returned.
1494    
1495    6.  If a subpattern that was called recursively or as a subroutine contained
1496        (*PRUNE) or any other control that caused it to give a non-standard return,
1497        invalid errors such as "Error -26 (nested recursion at the same subject
1498        position)" or even infinite loops could occur.
1499    
1500    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1501        computing the minimum length on reaching *ACCEPT, and so ended up with the
1502        wrong value of 1 rather than 0. Further investigation indicates that
1503        computing a minimum subject length in the presence of *ACCEPT is difficult
1504        (think back references, subroutine calls), and so I have changed the code
1505        so that no minimum is registered for a pattern that contains *ACCEPT.
1506    
1507    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1508        it was not handled as intended. [But see 16 below.]
1509    
1510    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1511        Sheri Pierce.
1512    
1513    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1514        the first byte in a match must be "a".
1515    
1516    11. Change 17 for 8.13 increased the recursion depth for patterns like
1517        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1518        pattern contains any instances of (*THEN). If it does not, the old
1519        optimizations are restored. It would be nice to do this on a per-group
1520        basis, but at the moment that is not feasible.
1521    
1522    12. In some environments, the output of pcretest -C is CRLF terminated. This
1523        broke RunTest's code that checks for the link size. A single white space
1524        character after the value is now allowed for.
1525    
1526    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1527        For "fr", it uses the Windows-specific input and output files.
1528    
1529    14. If (*THEN) appeared in a group that was called recursively or as a
1530        subroutine, it did not work as intended. [But see next item.]
1531    
1532    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1533        pattern fragments (but not containing any | characters). If A and B are
1534        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1535        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1536        D. In other words, Perl considers parentheses that do not contain any |
1537        characters to be part of a surrounding alternative, whereas PCRE was
1538        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1539        differently. PCRE now behaves in the same way as Perl, except in the case
1540        of subroutine/recursion calls such as (?1) which have in any case always
1541        been different (but PCRE had them first :-).
1542    
1543    16. Related to 15 above: Perl does not treat the | in a conditional group as
1544        creating alternatives. Such a group is treated in the same way as an
1545        ordinary group without any | characters when processing (*THEN). PCRE has
1546        been changed to match Perl's behaviour.
1547    
1548    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1549        RunGrepTest script failed.
1550    
1551    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1552        inevitable for groups that contain captures, but it can lead to a lot of
1553        stack use in large patterns. The old behaviour has been restored for atomic
1554        groups that do not contain any capturing parentheses.
1555    
1556    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1557        suppress the check for a minimum subject length at run time. (If it was
1558        given to pcre_exec() or pcre_dfa_exec() it did work.)
1559    
1560    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1561        fail to work when decoding hex characters in data strings in EBCDIC
1562        environments.
1563    
1564    21. It appears that in at least one Mac OS environment, the isxdigit() function
1565        is implemented as a macro that evaluates to its argument more than once,
1566        contravening the C 90 Standard (I haven't checked a later standard). There
1567        was an instance in pcretest which caused it to go wrong when processing
1568        \x{...} escapes in subject strings. The has been rewritten to avoid using
1569        things like p++ in the argument of isxdigit().
1570    
1571    
1572    Version 8.13 16-Aug-2011
1573    ------------------------
1574    
1575    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1576    
1577    2.  Two minor typos in pcre_internal.h have been fixed.
1578    
1579    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1580        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1581        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1582    
1583    4.  There were a number of related bugs in the code for matching backrefences
1584        caselessly in UTF-8 mode when codes for the characters concerned were
1585        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1586        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1587        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1588        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1589        2-byte code at the end of the subject (it thought there wasn't enough data
1590        left).
1591    
1592    5.  Comprehensive information about what went wrong is now returned by
1593        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1594        as the output vector has at least 2 elements. The offset of the start of
1595        the failing character and a reason code are placed in the vector.
1596    
1597    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1598        now returned is for the first byte of the failing character, instead of the
1599        last byte inspected. This is an incompatible change, but I hope it is small
1600        enough not to be a problem. It makes the returned offset consistent with
1601        pcre_exec() and pcre_dfa_exec().
1602    
1603    7.  pcretest now gives a text phrase as well as the error number when
1604        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1605        failure, the offset and reason code are output.
1606    
1607    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1608        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1609        back over a single character (\n). This seems wrong (because it treated the
1610        two characters as a single entity when going forwards), conflicts with the
1611        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1612        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1613        has been changed.
1614    
1615    9.  Some internal refactoring has changed the processing so that the handling
1616        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1617        time (the PCRE_DOTALL option was changed this way some time ago: version
1618        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1619        which was always a bit of a fudge. It also means that there is one less
1620        argument for the match() function, which reduces its stack requirements
1621        slightly. This change also fixes an incompatibility with Perl: the pattern
1622        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1623    
1624    10. More internal refactoring has drastically reduced the number of recursive
1625        calls to match() for possessively repeated groups such as (abc)++ when
1626        using pcre_exec().
1627    
1628    11. While implementing 10, a number of bugs in the handling of groups were
1629        discovered and fixed:
1630    
1631        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1632        (a|)*(?1) gave a compile-time internal error.
1633        ((a|)+)+  did not notice that the outer group could match an empty string.
1634        (^a|^)+   was not marked as anchored.
1635        (.*a|.*)+ was not marked as matching at start or after a newline.
1636    
1637    12. Yet more internal refactoring has removed another argument from the match()
1638        function. Special calls to this function are now indicated by setting a
1639        value in a variable in the "match data" data block.
1640    
1641    13. Be more explicit in pcre_study() instead of relying on "default" for
1642        opcodes that mean there is no starting character; this means that when new
1643        ones are added and accidentally left out of pcre_study(), testing should
1644        pick them up.
1645    
1646    14. The -s option of pcretest has been documented for ages as being an old
1647        synonym of -m (show memory usage). I have changed it to mean "force study
1648        for every regex", that is, assume /S for every regex. This is similar to -i
1649        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1650        using it. It makes it easier to run collections of tests with and without
1651        study enabled, and thereby test pcre_study() more easily. All the standard
1652        tests are now run with and without -s (but some patterns can be marked as
1653        "never study" - see 20 below).
1654    
1655    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1656        restoration of the capturing data to the outer values was not happening
1657        correctly.
1658    
1659    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1660        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1661        pattern had matched an empty string, and so incorrectly returned a no
1662        match.
1663    
1664    17. There was optimizing code for the last branch of non-capturing parentheses,
1665        and also for the obeyed branch of a conditional subexpression, which used
1666        tail recursion to cut down on stack usage. Unfortunately, now that there is
1667        the possibility of (*THEN) occurring in these branches, tail recursion is
1668        no longer possible because the return has to be checked for (*THEN). These
1669        two optimizations have therefore been removed. [But see 8.20/11 above.]
1670    
1671    18. If a pattern containing \R was studied, it was assumed that \R always
1672        matched two bytes, thus causing the minimum subject length to be
1673        incorrectly computed because \R can also match just one byte.
1674    
1675    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1676        was incorrectly computed.
1677    
1678    20. If /S is present twice on a test pattern in pcretest input, it now
1679        *disables* studying, thereby overriding the use of -s on the command line
1680        (see 14 above). This is necessary for one or two tests to keep the output
1681        identical in both cases.
1682    
1683    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1684        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1685    
1686    22. When an atomic group that contained a capturing parenthesis was
1687        successfully matched, but the branch in which it appeared failed, the
1688        capturing was not being forgotten if a higher numbered group was later
1689        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1690        group 1 to "a", when in fact it should be unset. This applied to multi-
1691        branched capturing and non-capturing groups, repeated or not, and also to
1692        positive assertions (capturing in negative assertions does not happen
1693        in PCRE) and also to nested atomic groups.
1694    
1695    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1696        subject after a captured substring, to make it easier to tell which of a
1697        number of identical substrings has been captured.
1698    
1699    24. The way atomic groups are processed by pcre_exec() has been changed so that
1700        if they are repeated, backtracking one repetition now resets captured
1701        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1702        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1703        "aaa". Previously, it would have been "a". As part of this code
1704        refactoring, the way recursive calls are handled has also been changed.
1705    
1706    25. If an assertion condition captured any substrings, they were not passed
1707        back unless some other capturing happened later. For example, if
1708        (?(?=(a))a) was matched against "a", no capturing was returned.
1709    
1710    26. When studying a pattern that contained subroutine calls or assertions,
1711        the code for finding the minimum length of a possible match was handling
1712        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1713        group 1 called group 2 while simultaneously a separate group 2 called group
1714        1). A stack overflow occurred in this case. I have fixed this by limiting
1715        the recursion depth to 10.
1716    
1717    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1718        Fortmann. This supports explicit test numbers on the command line, and has
1719        argument validation and error reporting.
1720    
1721    28. An instance of \X with an unlimited repeat could fail if at any point the
1722        first character it looked at was a mark character.
1723    
1724    29. Some minor code refactoring concerning Unicode properties and scripts
1725        should reduce the stack requirement of match() slightly.
1726    
1727    30. Added the '=' option to pcretest to check the setting of unused capturing
1728        slots at the end of the pattern, which are documented as being -1, but are
1729        not included in the return count.
1730    
1731    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1732        compiled something random. Now it gives a compile-time error (as does
1733        Perl).
1734    
1735    32. A *MARK encountered during the processing of a positive assertion is now
1736        recorded and passed back (compatible with Perl).
1737    
1738    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1739        had alternative anchored branches, the search for a second match in a line
1740        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1741        matched the line "0102" twice. The same bug affected patterns that started
1742        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1743        twice.
1744    
1745    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1746        does, and because of capturing effects, quantifying parenthesized
1747        assertions may at times be useful. Quantifiers are now allowed for
1748        parenthesized assertions.
1749    
1750    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1751    
1752    36. \g was being checked for fancy things in a character class, when it should
1753        just be a literal "g".
1754    
1755    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1756        appearance of a nested POSIX class supersedes an apparent external class.
1757        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1758        unescaped square brackets may also appear as part of class names. For
1759        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1760        more like Perl. (But see 8.20/1 above.)
1761    
1762    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1763        was because it thought it was \N{name}, which is not supported).
1764    
1765    39. Add minix to OS list not supporting the -S option in pcretest.
1766    
1767    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1768        cannot analyze patterns in sufficient detail to catch mutual recursions
1769        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1770        subgroup is called recursively as a subpattern for a second time at the
1771        same position in the subject string. In previous releases this might have
1772        been caught by the recursion limit, or it might have run out of stack.
1773    
1774    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1775        happen only once. PCRE was, however incorrectly giving a compile time error
1776        "recursive call could loop indefinitely" because it cannot analyze the
1777        pattern in sufficient detail. The compile time test no longer happens when
1778        PCRE is compiling a conditional subpattern, but actual runaway loops are
1779        now caught at runtime (see 40 above).
1780    
1781    42. It seems that Perl allows any characters other than a closing parenthesis
1782        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1783        has been changed to be the same.
1784    
1785    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1786        as not to get warnings when autogen.sh is called. Also changed
1787        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1788    
1789    44. To help people who use pcregrep to scan files containing exceedingly long
1790        lines, the following changes have been made:
1791    
1792        (a) The default value of the buffer size parameter has been increased from
1793            8K to 20K. (The actual buffer used is three times this size.)
1794    
1795        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1796            PCRE is built.
1797    
1798        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1799            to be set at run time.
1800    
1801        (d) Numerical values in pcregrep options can be followed by K or M, for
1802            example --buffer-size=50K.
1803    
1804        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1805            given and the return code is set to 2.
1806    
1807    45. Add a pointer to the latest mark to the callout data block.
1808    
1809    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1810        partial match of an empty string instead of no match. This was specific to
1811        the use of ".".
1812    
1813    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1814        complete match instead of a partial match. This bug was dependent on both
1815        the PCRE_UTF8 and PCRE_DOTALL options being set.
1816    
1817    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1818        starting byte set, because \b was not being ignored.
1819    
1820    
1821    Version 8.12 15-Jan-2011
1822    ------------------------
1823    
1824    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1825        checks for such things as part of the documentation building process.
1826    
1827    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1828        --match-limit and --recursion-limit options (added for 8.11). In
1829        particular, this made one of the standard tests fail. (The integer value
1830        went into the wrong half of a long int.)
1831    
1832    3.  If the --colour option was given to pcregrep with -v (invert match), it
1833        did strange things, either producing crazy output, or crashing. It should,
1834        of course, ignore a request for colour when reporting lines that do not
1835        match.
1836    
1837    4.  Another pcregrep bug caused similar problems if --colour was specified with
1838        -M (multiline) and the pattern match finished with a line ending.
1839    
1840    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1841        matched in multiline mode, the following line was shown as part of the
1842        match. This seems wrong, so I have changed it.
1843    
1844    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1845        the check for further matches in the same line (so they could be coloured)
1846        to overrun the end of the current line. If another match was found, it was
1847        incorrectly shown (and then shown again when found in the next line).
1848    
1849    7.  If pcregrep was compiled under Windows, there was a reference to the
1850        function pcregrep_exit() before it was defined. I am assuming this was
1851        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1852        reported by a user. I've moved the definition above the reference.
1853    
1854    
1855    Version 8.11 10-Dec-2010
1856    ------------------------
1857    
1858    1.  (*THEN) was not working properly if there were untried alternatives prior
1859        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1860        backtracked to try for "b" instead of moving to the next alternative branch
1861        at the same level (in this case, to look for "c"). The Perl documentation
1862        is clear that when (*THEN) is backtracked onto, it goes to the "next
1863        alternative in the innermost enclosing group".
1864    
1865    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1866        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1867        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1868        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1869        (*THEN).
1870    
1871    3.  If \s appeared in a character class, it removed the VT character from
1872        the class, even if it had been included by some previous item, for example
1873        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1874        of \s, but is part of the POSIX "space" class.)
1875    
1876    4.  A partial match never returns an empty string (because you can always
1877        match an empty string at the end of the subject); however the checking for
1878        an empty string was starting at the "start of match" point. This has been
1879        changed to the "earliest inspected character" point, because the returned
1880        data for a partial match starts at this character. This means that, for
1881        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1882        (previously it gave "no match").
1883    
1884    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1885        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1886        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1887        has an implication that the given string is incomplete (because a partial
1888        match is preferred over a full match). For this reason, these items now
1889        give a partial match in this situation. [Aside: previously, the one case
1890        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1891        match rather than a full match, which was wrong by the old rules, but is
1892        now correct.]
1893    
1894    6.  There was a bug in the handling of #-introduced comments, recognized when
1895        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1896        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1897        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1898        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1899        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1900        places in pcre_compile().
1901    
1902    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1903        comments when looking ahead for named forward references to subpatterns,
1904        the only newline sequence it recognized was NL. It now handles newlines
1905        according to the set newline convention.
1906    
1907    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1908        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1909        cater for a lack of strerror(). These oversights have been fixed.
1910    
1911    9.  Added --match-limit and --recursion-limit to pcregrep.
1912    
1913    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1914    
1915    11. When the -o option was used, pcregrep was setting a return code of 1, even
1916        when matches were found, and --line-buffered was not being honoured.
1917    
1918    12. Added an optional parentheses number to the -o and --only-matching options
1919        of pcregrep.
1920    
1921    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1922        can match an empty string. The code to do it in pcretest and pcredemo
1923        needed fixing:
1924    
1925        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1926            only one byte after an empty string match just before CRLF (this case
1927            just got forgotten; "any" and "anycrlf" were OK).
1928    
1929        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1930            mode when an empty string match preceded an ASCII character followed by
1931            a non-ASCII character. (The code for advancing by one character rather
1932            than one byte was nonsense.)
1933    
1934        (c) The pcredemo.c sample program did not have any code at all to handle
1935            the cases when CRLF is a valid newline sequence.
1936    
1937    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1938        as a starting offset was within the subject string. There is now a new
1939        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1940        negative or greater than the length of the string. In order to test this,
1941        pcretest is extended to allow the setting of negative starting offsets.
1942    
1943    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1944        starting offset points to the beginning of a UTF-8 character was
1945        unnecessarily clumsy. I tidied it up.
1946    
1947    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1948        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1949    
1950    17. Nobody had reported that the --include_dir option, which was added in
1951        release 7.7 should have been called --include-dir (hyphen, not underscore)
1952        for compatibility with GNU grep. I have changed it to --include-dir, but
1953        left --include_dir as an undocumented synonym, and the same for
1954        --exclude-dir, though that is not available in GNU grep, at least as of
1955        release 2.5.4.
1956    
1957    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1958        characters from a string of bytes) have been redefined so as not to use
1959        loops, in order to improve performance in some environments. At the same
1960        time, I abstracted some of the common code into auxiliary macros to save
1961        repetition (this should not affect the compiled code).
1962    
1963    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1964        compile-time error is now given if \c is not followed by an ASCII
1965        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1966        different, and any byte value is allowed.)
1967    
1968    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1969        START_OPTIMIZE option, which is now allowed at compile time - but just
1970        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1971        to pcregrep and other applications that have no direct access to PCRE
1972        options. The new /Y option in pcretest sets this option when calling
1973        pcre_compile().
1974    
1975    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1976        back references. Groups containing recursive back references were forced to
1977        be atomic by that change, but in the case of named groups, the amount of
1978        memory required was incorrectly computed, leading to "Failed: internal
1979        error: code overflow". This has been fixed.
1980    
1981    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1982        pcretest.c, to avoid build problems in some Borland environments.
1983    
1984    
1985    Version 8.10 25-Jun-2010
1986    ------------------------
1987    
1988    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1989        THEN.
1990    
1991    2.  (*ACCEPT) was not working when inside an atomic group.
1992    
1993    3.  Inside a character class, \B is treated as a literal by default, but
1994        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1995        causes the error). The code is unchanged, but I tidied the documentation.
1996    
1997    4.  Inside a character class, PCRE always treated \R and \X as literals,
1998        whereas Perl faults them if its -w option is set. I have changed PCRE so
1999        that it faults them when PCRE_EXTRA is set.
2000    
2001    5.  Added support for \N, which always matches any character other than
2002        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2003    
2004    6.  When compiling pcregrep with newer versions of gcc which may have
2005        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2006        declared with attribute warn_unused_result" were given. Just casting the
2007        result to (void) does not stop the warnings; a more elaborate fudge is
2008        needed. I've used a macro to implement this.
2009    
2010    7.  Minor change to pcretest.c to avoid a compiler warning.
2011    
2012    8.  Added four artifical Unicode properties to help with an option to make
2013        \s etc use properties (see next item). The new properties are: Xan
2014        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2015    
2016    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2017        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2018        this option. Modified pcretest to add /W to test this facility. Added
2019        REG_UCP to make it available via the POSIX interface.
2020    
2021    10. Added --line-buffered to pcregrep.
2022    
2023    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2024        studied, and the match started with a letter with a code point greater than
2025        127 whose first byte was different to the first byte of the other case of
2026        the letter, the other case of this starting letter was not recognized
2027        (#976).
2028    
2029    12. If a pattern that was studied started with a repeated Unicode property
2030        test, for example, \p{Nd}+, there was the theoretical possibility of
2031        setting up an incorrect bitmap of starting bytes, but fortunately it could
2032        not have actually happened in practice until change 8 above was made (it
2033        added property types that matched character-matching opcodes).
2034    
2035    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2036        possible starting bytes for non-anchored patterns.
2037    
2038    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2039        \R, and also a number of cases that involve Unicode properties, both
2040        explicit and implicit when PCRE_UCP is set.
2041    
2042    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2043        input, it could crash or give wrong results if characters with values
2044        greater than 0xc0 were present in the subject string. (Detail: it assumed
2045        UTF-8 input when processing these items.)
2046    
2047    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2048        size_t is 64-bit (#991).
2049    
2050    17. Added a check for running out of memory when PCRE is compiled with
2051        --disable-stack-for-recursion (#990).
2052    
2053    18. If the last data line in a file for pcretest does not have a newline on
2054        the end, a newline was missing in the output.
2055    
2056    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2057        less than 128) in its various bitmaps. However, there is a facility for
2058        generating tables according to the current locale when PCRE is compiled. It
2059        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2060        characters, are recognized by isspace() and therefore were getting set in
2061        these tables, and indeed these tables seem to approximate to ISO 8859. This
2062        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2063        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2064        which of course cannot start UTF-8 characters. I have changed the code so
2065        that only real ASCII characters (less than 128) and the correct starting
2066        bytes for UTF-8 encodings are set for characters greater than 127 when in
2067        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2068        altogether.)
2069    
2070    20. Added the /T option to pcretest so as to be able to run tests with non-
2071        standard character tables, thus making it possible to include the tests
2072        used for 19 above in the standard set of tests.
2073    
2074    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2075        reference to a subpattern the other side of a comment that contains an
2076        opening parenthesis caused either an internal compiling error, or a
2077        reference to the wrong subpattern.
2078    
2079    
2080    Version 8.02 19-Mar-2010
2081    ------------------------
2082    
2083    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2084    
2085    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2086        configured.
2087    
2088    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2089        original author of that file, following a query about its status.
2090    
2091    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2092        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2093    
2094    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2095        quantifier applied to a forward-referencing subroutine call, could compile
2096        incorrect code or give the error "internal error: previously-checked
2097        referenced subpattern not found".
2098    
2099    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2100        variables to point to external functions. For these systems, therefore,
2101        pcre_malloc etc. are now initialized to local functions that call the
2102        relevant global functions.
2103    
2104    7.  There were two entries missing in the vectors called coptable and poptable
2105        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2106        I've fixed the data, and added a kludgy way of testing at compile time that
2107        the lengths are correct (equal to the number of opcodes).
2108    
2109    8.  Following on from 7, I added a similar kludge to check the length of the
2110        eint vector in pcreposix.c.
2111    
2112    9.  Error texts for pcre_compile() are held as one long string to avoid too
2113        much relocation at load time. To find a text, the string is searched,
2114        counting zeros. There was no check for running off the end of the string,
2115        which could happen if a new error number was added without updating the
2116        string.
2117    
2118    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2119    
2120    11. \K was not working if it appeared in an atomic group or in a group that
2121        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2122        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2123        the assertion is positive, but not if it is negative.
2124    
2125    12. Change 11 fortuitously reduced the size of the stack frame used in the
2126        "match()" function of pcre_exec.c by one pointer. Forthcoming
2127        implementation of support for (*MARK) will need an extra pointer on the
2128        stack; I have reserved it now, so that the stack frame size does not
2129        decrease.
2130    
2131    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2132        item in branch that calls a recursion is a subroutine call - as in the
2133        second branch in the above example - was incorrectly given the compile-
2134        time error "recursive call could loop indefinitely" because pcre_compile()
2135        was not correctly checking the subroutine for matching a non-empty string.
2136    
2137    14. The checks for overrunning compiling workspace could trigger after an
2138        overrun had occurred. This is a "should never occur" error, but it can be
2139        triggered by pathological patterns such as hundreds of nested parentheses.
2140        The checks now trigger 100 bytes before the end of the workspace.
2141    
2142    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2143    
2144    
2145    Version 8.01 19-Jan-2010
2146    ------------------------
2147    
2148    1.  If a pattern contained a conditional subpattern with only one branch (in
2149        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2150        computed the wrong minimum data length (which is of course zero for such
2151        subpatterns). This could cause incorrect "no match" results.
2152    
2153    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2154        the pattern is reset in the first branch, pcre_compile() failed with
2155        "internal error: code overflow at offset...". This happened only when
2156        the reset was to the original external option setting. (An optimization
2157        abstracts leading options settings into an external setting, which was the
2158        cause of this.)
2159    
2160    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2161        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2162        assertion pattern did not match (meaning that the assertion was true), it
2163        was incorrectly treated as false if the SKIP had been reached during the
2164        matching. This also applied to assertions used as conditions.
2165    
2166    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2167        assertion subpattern, including such a pattern used as a condition,
2168        unpredictable results occurred, instead of the error return
2169        PCRE_ERROR_DFA_UITEM.
2170    
2171    5.  The C++ GlobalReplace function was not working like Perl for the special
2172        situation when an empty string is matched. It now does the fancy magic
2173        stuff that is necessary.
2174    
2175    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2176        removed. (These were left over from very, very early versions of PCRE.)
2177    
2178    7.  Some cosmetic changes to the code to make life easier when compiling it
2179        as part of something else:
2180    
2181        (a) Change DEBUG to PCRE_DEBUG.
2182    
2183        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2184            called "current" as "current_branch", to prevent a collision with the
2185            Linux macro when compiled as a kernel module.
2186    
2187        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2188            prevent a collision with the Linux macro when compiled as a kernel
2189            module.
2190    
2191    8.  In pcre_compile() there are some checks for integer overflows that used to
2192        cast potentially large values to (double). This has been changed to that
2193        when building, a check for int64_t is made, and if it is found, it is used
2194        instead, thus avoiding the use of floating point arithmetic. (There is no
2195        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2196        double.
2197    
2198    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2199        2005 (difference between two addresses compared to an unsigned value).
2200    
2201    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2202        custom one, because of the following reported problem in Windows:
2203    
2204          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2205              under Win32.
2206          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2207              therefore missing the function definition.
2208          - The compiler thus generates a "C" signature for the test function.
2209          - The linker fails to find the "C" function.
2210          - PCRE fails to configure if asked to do so against libbz2.
2211    
2212    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2213        messages were output:
2214    
2215          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2216          rerunning libtoolize, to keep the correct libtool macros in-tree.
2217          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2218    
2219        I have done both of these things.
2220    
2221    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2222        most of the time, it *can* run out if it is given a pattern that contains a
2223        runaway infinite recursion. I updated the discussion in the pcrestack man
2224        page.
2225    
2226    13. Now that we have gone to the x.xx style of version numbers, the minor
2227        version may start with zero. Using 08 or 09 is a bad idea because users
2228        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2229        interpreted as invalid octal numbers. I've updated the previous comment in
2230        configure.ac, and also added a check that gives an error if 08 or 09 are
2231        used.
2232    
2233    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2234        causing partial matching to fail when the end of the subject matched \W
2235        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2236    
2237    15. There were some discrepancies between the declarations in pcre_internal.h
2238        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2239        their definitions. The declarations used "const uschar *" and the
2240        definitions used USPTR. Even though USPTR is normally defined as "const
2241        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2242        reported that: "This difference in casting confuses some C++ compilers, for
2243        example, SunCC recognizes above declarations as different functions and
2244        generates broken code for hbpcre." I have changed the declarations to use
2245        USPTR.
2246    
2247    16. GNU libtool is named differently on some systems. The autogen.sh script now
2248        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2249        (FreeBSD).
2250    
2251    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2252        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2253        comment: "Figure out how to create a longlong from a string: strtoll and
2254        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2255        instance, but it only takes 2 args instead of 3!"
2256    
2257    18. A subtle bug concerned with back references has been fixed by a change of
2258        specification, with a corresponding code fix. A pattern such as
2259        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2260        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2261        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2262        same bug. Such groups have to be quantified to be useful, or contained
2263        inside another quantified group. (If there's no repetition, the reference
2264        can never match.) The problem arises because, having left the group and
2265        moved on to the rest of the pattern, a later failure that backtracks into
2266        the group uses the captured value from the final iteration of the group
2267        rather than the correct earlier one. I have fixed this in PCRE by forcing
2268        any group that contains a reference to itself to be an atomic group; that
2269        is, there cannot be any backtracking into it once it has completed. This is
2270        similar to recursive and subroutine calls.
2271    
2272    
2273    Version 8.00 19-Oct-09
2274    ----------------------
2275    
2276    1.  The table for translating pcre_compile() error codes into POSIX error codes
2277        was out-of-date, and there was no check on the pcre_compile() error code
2278        being within the table. This could lead to an OK return being given in
2279        error.
2280    
2281    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2282        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2283        in a Windows environment.
2284    
2285    3.  The pcregrep --count option prints the count for each file even when it is
2286        zero, as does GNU grep. However, pcregrep was also printing all files when
2287        --files-with-matches was added. Now, when both options are given, it prints
2288        counts only for those files that have at least one match. (GNU grep just
2289        prints the file name in this circumstance, but including the count seems
2290        more useful - otherwise, why use --count?) Also ensured that the
2291        combination -clh just lists non-zero counts, with no names.
2292    
2293    4.  The long form of the pcregrep -F option was incorrectly implemented as
2294        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2295        but it seems right to fix it, and I didn't think it was worth preserving
2296        the old behaviour.
2297    
2298    5.  The command line items --regex=pattern and --regexp=pattern were not
2299        recognized by pcregrep, which required --regex pattern or --regexp pattern
2300        (with a space rather than an '='). The man page documented the '=' forms,
2301        which are compatible with GNU grep; these now work.
2302    
2303    6.  No libpcreposix.pc file was created for pkg-config; there was just
2304        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2305    
2306    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2307        when UCP support is not needed, by modifying the Python script that
2308        generates it from Unicode data files. This should not matter if the module
2309        is correctly used as a library, but I received one complaint about 50K of
2310        unwanted data. My guess is that the person linked everything into his
2311        program rather than using a library. Anyway, it does no harm.
2312    
2313    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2314        was a minimum greater than 1 for a wide character in a possessive
2315        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2316        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2317        character. Chaos in the form of incorrect output or a compiling loop could
2318        result.
2319    
2320    9.  The restrictions on what a pattern can contain when partial matching is
2321        requested for pcre_exec() have been removed. All patterns can now be
2322        partially matched by this function. In addition, if there are at least two
2323        slots in the offset vector, the offset of the earliest inspected character
2324        for the match and the offset of the end of the subject are set in them when
2325        PCRE_ERROR_PARTIAL is returned.
2326    
2327    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2328        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2329        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2330        and may be more useful for multi-segment matching.
2331    
2332    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2333        used to be given if ever the end of the subject was reached; now it is
2334        given only if matching could not proceed because another character was
2335        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2336        string "Z", which now yields "no match" instead of "partial match". In the
2337        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2338        final character ended with (*FAIL).
2339    
2340    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2341        if the pattern had a "must contain" character that was already found in the
2342        earlier partial match, unless partial matching was again requested. For
2343        example, with the pattern /dog.(body)?/, the "must contain" character is
2344        "g". If the first part-match was for the string "dog", restarting with
2345        "sbody" failed. This bug has been fixed.
2346    
2347    13. The string returned by pcre_dfa_exec() after a partial match has been
2348        changed so that it starts at the first inspected character rather than the
2349        first character of the match. This makes a difference only if the pattern
2350        starts with a lookbehind assertion or \b or \B (\K is not supported by
2351        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2352        matching functions compatible, and I think it's the right thing to do.
2353    
2354    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2355        so that the demonstration program is easily available in environments where
2356        PCRE has not been installed from source.
2357    
2358    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2359        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2360        library.
2361    
2362    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2363        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2364        is not the first non-POSIX option to be added. Clearly some people find
2365        these options useful.
2366    
2367    17. If a caller to the POSIX matching function regexec() passes a non-zero
2368        value for nmatch with a NULL value for pmatch, the value of
2369        nmatch is forced to zero.
2370    
2371    18. RunGrepTest did not have a test for the availability of the -u option of
2372        the diff command, as RunTest does. It now checks in the same way as
2373        RunTest, and also checks for the -b option.
2374    
2375    19. If an odd number of negated classes containing just a single character
2376        interposed, within parentheses, between a forward reference to a named
2377        subpattern and the definition of the subpattern, compilation crashed with
2378        an internal error, complaining that it could not find the referenced
2379        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2380        [The bug was that it was starting one character too far in when skipping
2381        over the character class, thus treating the ] as data rather than
2382        terminating the class. This meant it could skip too much.]
2383    
2384    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2385        /g option in pcretest when the pattern contains \K, which makes it possible
2386        to have an empty string match not at the start, even when the pattern is
2387        anchored. Updated pcretest and pcredemo to use this option.
2388    
2389    21. If the maximum number of capturing subpatterns in a recursion was greater
2390        than the maximum at the outer level, the higher number was returned, but
2391        with unset values at the outer level. The correct (outer level) value is
2392        now given.
2393    
2394    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2395        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2396        make it do so. The string so far is captured, making this feature
2397        compatible with Perl.
2398    
2399    23. The tests have been re-organized, adding tests 11 and 12, to make it
2400        possible to check the Perl 5.10 features against Perl 5.10.
2401    
2402    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2403        pattern matches a fixed length string. PCRE did not allow this; now it
2404        does. Neither allows recursion.
2405    
2406    25. I finally figured out how to implement a request to provide the minimum
2407        length of subject string that was needed in order to match a given pattern.
2408        (It was back references and recursion that I had previously got hung up
2409        on.) This code has now been added to pcre_study(); it finds a lower bound
2410        to the length of subject needed. It is not necessarily the greatest lower
2411        bound, but using it to avoid searching strings that are too short does give
2412        some useful speed-ups. The value is available to calling programs via
2413        pcre_fullinfo().
2414    
2415    26. While implementing 25, I discovered to my embarrassment that pcretest had
2416        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2417        study optimizations had never been tested with that matching function.
2418        Oops. What is worse, even when it was passed study data, there was a bug in
2419        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2420        were also very few tests of studied patterns with pcre_dfa_exec().
2421    
2422    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2423        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2424        on the other side of the coin, they are no longer allowed to have different
2425        names, because these cannot be distinguished in PCRE, and this has caused
2426        confusion. (This is a difference from Perl.)
2427    
2428    28. When duplicate subpattern names are present (necessarily with different
2429        numbers, as required by 27 above), and a test is made by name in a
2430        conditional pattern, either for a subpattern having been matched, or for
2431        recursion in such a pattern, all the associated numbered subpatterns are
2432        tested, and the overall condition is true if the condition is true for any
2433        one of them. This is the way Perl works, and is also more like the way
2434        testing by number works.
2435    
2436    
2437    Version 7.9 11-Apr-09
2438    ---------------------
2439    
2440    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2441        (pcretest), all targets were linked against these libraries. This included
2442        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2443        libraries. This caused unwanted dependencies to be created. This problem
2444        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2445        pcretest is linked with readline.
2446    
2447    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2448        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2449        moved outside it again, because FALSE and TRUE are already defined in AIX,
2450        but BOOL is not.
2451    
2452    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2453        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2454    
2455    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2456        hyphens) following file names and line numbers when outputting matching
2457        lines. This is not true; no spaces are inserted. I have also clarified the
2458        wording for the --colour (or --color) option.
2459    
2460    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2461        was not coloured; this is different to GNU grep, so I have changed it to be
2462        the same.
2463    
2464    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2465        each matching line was coloured. Now it goes on to look for further matches
2466        of any of the test patterns, which is the same behaviour as GNU grep.
2467    
2468    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2469        doesn't make sense to accept an empty string match in pcregrep, so I have
2470        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2471        seems to be how GNU grep behaves. [But see later change 40 for release
2472        8.33.]
2473    
2474    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2475        start or after a newline", because the conditional assertion was not being
2476        correctly handled. The rule now is that both the assertion and what follows
2477        in the first alternative must satisfy the test.
2478    
2479    9.  If auto-callout was enabled in a pattern with a conditional group whose
2480        condition was an assertion, PCRE could crash during matching, both with
2481        pcre_exec() and pcre_dfa_exec().
2482    
2483    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2484        used for matching.
2485    
2486    11. Unicode property support in character classes was not working for
2487        characters (bytes) greater than 127 when not in UTF-8 mode.
2488    
2489    12. Added the -M command line option to pcretest.
2490    
2491    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2492    
2493    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2494    
2495    16. Added comments and documentation about mis-use of no_arg in the C++
2496        wrapper.
2497    
2498    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2499        from Martin Jerabek that uses macro names for all relevant character and
2500        string constants.
2501    
2502    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2503        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2504        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2505        these, but not everybody uses configure.
2506    
2507    19. A conditional group that had only one branch was not being correctly
2508        recognized as an item that could match an empty string. This meant that an
2509        enclosing group might also not be so recognized, causing infinite looping
2510        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2511        with the subject "ab", where knowledge that the repeated group can match
2512        nothing is needed in order to break the loop.
2513    
2514    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2515        exec(), but without supplying a callout function, matching went wrong.
2516    
2517    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2518        leak if the size of the offset vector was greater than 30. When the vector
2519        is smaller, the saved offsets during recursion go onto a local stack
2520        vector, but for larger vectors malloc() is used. It was failing to free
2521        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2522        error, in fact).
2523    
2524    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2525        heapframe that is used only when UTF-8 support is enabled. This caused no
2526        problem, but was untidy.
2527    
2528    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2529        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2530        included within another project.
2531    
2532    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2533        slightly modified by me:
2534    
2535          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2536              not building pcregrep.
2537    
2538          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2539              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2540    
2541    25. Forward references, both numeric and by name, in patterns that made use of
2542        duplicate group numbers, could behave incorrectly or give incorrect errors,
2543        because when scanning forward to find the reference group, PCRE was not
2544        taking into account the duplicate group numbers. A pattern such as
2545        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2546    
2547    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2548        the feature of a custom pointer more persuasive (as requested by a user).
2549    
2550    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2551        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2552        pre-defined.
2553    
2554    28. Added support for (*UTF8) at the start of a pattern.
2555    
2556    29. Arrange for flags added by the "release type" setting in CMake to be shown
2557        in the configuration summary.
2558    
2559    
2560    Version 7.8 05-Sep-08
2561    ---------------------
2562    
2563    1.  Replaced UCP searching code with optimized version as implemented for Ad
2564        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2565        stage table and inline lookup instead of a function, giving speed ups of 2
2566        to 5 times on some simple patterns that I tested. Permission was given to
2567        distribute the MultiStage2.py script that generates the tables (it's not in
2568        the tarball, but is in the Subversion repository).
2569    
2570    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2571        scripts.
2572    
2573    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2574        a group with a zero qualifier. The result of the study could be incorrect,
2575        or the function might crash, depending on the pattern.
2576    
2577    4.  Caseless matching was not working for non-ASCII characters in back
2578        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2579        It now works when Unicode Property Support is available.
2580    
2581    5.  In pcretest, an escape such as \x{de} in the data was always generating
2582        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2583        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2584        truncation.
2585    
2586    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2587    
2588    7.  Added two (int) casts to pcregrep when printing the difference of two
2589        pointers, in case they are 64-bit values.
2590    
2591    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2592        test 2 if it fails.
2593    
2594    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2595        and a #define of that name to empty if it is not externally set. This is to
2596        allow users of MSVC to set it if necessary.
2597    
2598    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2599        the convenience functions in the pcre_get.c source file.
2600    
2601    11. An option change at the start of a pattern that had top-level alternatives
2602        could cause overwriting and/or a crash. This command provoked a crash in
2603        some environments:
2604    
2605          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2606    
2607        This potential security problem was recorded as CVE-2008-2371.
2608    
2609    12. For a pattern where the match had to start at the beginning or immediately
2610        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2611        pcre_dfa_exec() could read past the end of the passed subject if there was
2612        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2613        pcretest so that it places the subject at the end of its malloc-ed buffer.
2614    
2615    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2616        exec() might read past the end of the data buffer in UTF-8 mode.
2617    
2618    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2619        the data contained the byte 0x85 as part of a UTF-8 character within its
2620        first line. This applied both to normal and DFA matching.
2621    
2622    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2623        /^[^d]*?$/8 failed to match "abc".
2624    
2625    16. Added a missing copyright notice to pcrecpp_internal.h.
2626    
2627    17. Make it more clear in the documentation that values returned from
2628        pcre_exec() in ovector are byte offsets, not character counts.
2629    
2630    18. Tidied a few places to stop certain compilers from issuing warnings.
2631    
2632    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2633        supplied by Stefan Weber. I made a further small update for 7.8 because
2634        there is a change of source arrangements: the pcre_searchfuncs.c module is
2635        replaced by pcre_ucd.c.
2636    
2637    
2638    Version 7.7 07-May-08
2639    ---------------------
2640    
2641    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2642        a string to a long long, pretend we don't even have a long long." This is
2643        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2644    
2645    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2646        pre-7.6 versions, which defined a global no_arg variable instead of putting
2647        it in the RE class. (See also #8 below.)
2648    
2649    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2650        Lopes.
2651    
2652    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2653    
2654        (1) The include/exclude patterns were being applied to the whole pathnames
2655            of files, instead of just to the final components.
2656    
2657        (2) If there was more than one level of directory, the subdirectories were
2658            skipped unless they satisfied the include/exclude conditions. This is
2659            inconsistent with GNU grep (and could even be seen as contrary to the
2660            pcregrep specification - which I improved to make it absolutely clear).
2661            The action now is always to scan all levels of directory, and just
2662            apply the include/exclude patterns to regular files.
2663    
2664    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2665        --exclude_dir in the tests to avoid scanning .svn directories.
2666    
2667    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2668        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2669        doesn't support NULs in patterns.
2670    
2671    7.  Added some missing "const"s to declarations of static tables in
2672        pcre_compile.c and pcre_dfa_exec.c.
2673    
2674    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2675        caused by fix #2  above. (Subsequently also a second patch to fix the
2676        first patch. And a third patch - this was a messy problem.)
2677    
2678    9.  Applied Craig's patch to remove the use of push_back().
2679    
2680    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2681        matching function regexec().
2682    
2683    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2684        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2685        references. PCRE supports relative numbers with this syntax (I don't think
2686        Oniguruma does).
2687    
2688    12. Previously, a group with a zero repeat such as (...){0} was completely
2689        omitted from the compiled regex. However, this means that if the group
2690        was called as a subroutine from elsewhere in the pattern, things went wrong
2691        (an internal error was given). Such groups are now left in the compiled
2692        pattern, with a new opcode that causes them to be skipped at execution
2693        time.
2694    
2695    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2696        to the way PCRE behaves:
2697    
2698        (a) A lone ] character is dis-allowed (Perl treats it as data).
2699    
2700        (b) A back reference to an unmatched subpattern matches an empty string
2701            (Perl fails the current match path).
2702    
2703        (c) A data ] in a character class must be notated as \] because if the
2704            first data character in a class is ], it defines an empty class. (In
2705            Perl it is not possible to have an empty class.) The empty class []
2706            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2707            The negative empty class [^] matches any one character, independently
2708            of the DOTALL setting.
2709    
2710    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2711        non-existent subpattern following a character class starting with ']' and
2712        containing () gave an internal compiling error instead of "reference to
2713        non-existent subpattern". Fortunately, when the pattern did exist, the
2714        compiled code was correct. (When scanning forwards to check for the
2715        existence of the subpattern, it was treating the data ']' as terminating
2716        the class, so got the count wrong. When actually compiling, the reference
2717        was subsequently set up correctly.)
2718    
2719    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2720        it was being rejected as not supported by pcre_dfa_exec(), even though
2721        other assertions are supported. I have made pcre_dfa_exec() support
2722        (*FAIL).
2723    
2724    16. The implementation of 13c above involved the invention of a new opcode,
2725        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2726        cannot be changed at match time, I realized I could make a small
2727        improvement to matching performance by compiling OP_ALLANY instead of
2728        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2729        on the OP_ANY path.
2730    
2731    17. Compiling pcretest on Windows with readline support failed without the
2732        following two fixes: (1) Make the unistd.h include conditional on
2733        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2734    
2735    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2736        ncurses library to be included for pcretest when ReadLine support is
2737        requested, but also to allow for it to be overridden. This patch came from
2738        Daniel Bergström.
2739    
2740    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2741        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2742        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2743        spotting this.
2744    
2745    
2746    Version 7.6 28-Jan-08
2747    ---------------------
2748    
2749    1.  A character class containing a very large number of characters with
2750        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2751        overflow.
2752    
2753    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2754        HAVE_LONG_LONG is not defined.
2755    
2756    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2757        bring them up to date and include new features. This patch includes:
2758    
2759        - Fixed PH's badly added libz and libbz2 support.
2760        - Fixed a problem with static linking.
2761        - Added pcredemo. [But later removed - see 7 below.]
2762        - Fixed dftables problem and added an option.
2763        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2764            HAVE_LONG_LONG.
2765        - Added readline support for pcretest.
2766        - Added an listing of the option settings after cmake has run.
2767    
2768    4.  A user submitted a patch to Makefile that makes it easy to create
2769        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2770        Makefile.am that cause it to include this special target, without
2771        affecting anything else. Note that the same mingw target plus all
2772        the other distribution libraries and programs are now supported
2773        when configuring with CMake (see 6 below) instead of with
2774        Configure/Make.
2775    
2776    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2777        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2778        exported in the Windows port". It has not yet been confirmed that the patch
2779        solves the problem, but it does no harm.
2780    
2781    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2782        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2783        with CMake, and also correct the comment about stack recursion.
2784    
2785    7.  Remove the automatic building of pcredemo from the ./configure system and
2786        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2787        of a program that users should build themselves after PCRE is installed, so
2788        building it automatically is not really right. What is more, it gave
2789        trouble in some build environments.
2790    
2791    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2792    
2793    
2794    Version 7.5 10-Jan-08
2795    ---------------------
2796    
2797    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2798        values in parens when parsing an RE using the C++ wrapper."
2799    
2800    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2801        Characters greater than 255 were excluded from the class instead of being
2802        included.
2803    
2804    3.  The same bug as (2) above applied to negated POSIX classes such as
2805        [:^space:].
2806    
2807    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2808        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2809        I have changed it.
2810    
2811    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2812        first named subpattern) and a construct such as (?&a) would reference the
2813        first named subpattern whose name started with "a" (in other words, the
2814        length check was missing). Both these problems are fixed. "Subpattern name
2815        expected" is now given for (?&) (a zero-length name), and this patch also
2816        makes it give the same error for \k'' (previously it complained that that
2817        was a reference to a non-existent subpattern).
2818    
2819    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2820        this is right because (?- can be followed by option settings as well as by
2821        digits. I have, however, made the messages clearer.
2822    
2823    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2824        than the number used in the conditional) now cause a compile-time error.
2825        This is actually not compatible with Perl, which accepts such patterns, but
2826        treats the conditional as always being FALSE (as PCRE used to), but it
2827        seems to me that giving a diagnostic is better.
2828    
2829    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2830        and messages.
2831    
2832    9.  Fix two occurrences of "backslash" in comments that should have been
2833        "backspace".
2834    
2835    10. Remove two redundant lines of code that can never be obeyed (their function
2836        was moved elsewhere).
2837    
2838    11. The program that makes PCRE's Unicode character property table had a bug
2839        which caused it to generate incorrect table entries for sequences of
2840        characters that have the same character type, but are in different scripts.
2841        It amalgamated them into a single range, with the script of the first of
2842        them. In other words, some characters were in the wrong script. There were
2843        thirteen such cases, affecting characters in the following ranges:
2844    
2845          U+002b0 - U+002c1
2846          U+0060c - U+0060d
2847          U+0061e - U+00612
2848          U+0064b - U+0065e
2849          U+0074d - U+0076d
2850          U+01800 - U+01805
2851          U+01d00 - U+01d77
2852          U+01d9b - U+01dbf
2853          U+0200b - U+0200f
2854          U+030fc - U+030fe
2855          U+03260 - U+0327f
2856          U+0fb46 - U+0fbb1
2857          U+10450 - U+1049d
2858    
2859    12. The -o option (show only the matching part of a line) for pcregrep was not
2860        compatible with GNU grep in that, if there was more than one match in a
2861        line, it showed only the first of them. It now behaves in the same way as
2862        GNU grep.
2863    
2864    13. If the -o and -v options were combined for pcregrep, it printed a blank
2865        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2866        does the same. The return code can be used to tell if there were any
2867        non-matching lines.
2868    
2869    14. Added --file-offsets and --line-offsets to pcregrep.
2870    
2871    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2872        infinitely looping recursion. The bug was that positive lookaheads were not
2873        being skipped when checking for a possible empty match (negative lookaheads
2874        and both kinds of lookbehind were skipped).
2875    
2876    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2877        inclusion of <windows.h> to before rather than after the definition of
2878        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2879    
2880    17. Specifying a possessive quantifier with a specific limit for a Unicode
2881        character property caused pcre_compile() to compile bad code, which led at
2882        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2883        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2884        caused the error; without that there was no problem.
2885    
2886    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2887    
2888    19. Added --enable-pcretest-libreadline.
2889    
2890    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2891        RE::GlobalReplace(). As a result, the number of replacements returned was
2892        double what it should be. I removed one of the increments, but Craig sent a
2893        later patch that removed the other one (the right fix) and added unit tests
2894        that check the return values (which was not done before).
2895    
2896    21. Several CMake things:
2897    
2898        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2899            the names libpcre and libpcreposix, not just pcre and pcreposix.
2900    
2901        (2) The above change means that pcretest and pcregrep are now correctly
2902            linked with the newly-built libraries, not previously installed ones.
2903    
2904        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2905    
2906    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2907        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2908        UTF-8 newline character). The key issue is that the pattern starts .*;
2909        this means that the match must be either at the beginning, or after a
2910        newline. The bug was in the code for advancing after a failed match and
2911        checking that the new position followed a newline. It was not taking
2912        account of UTF-8 characters correctly.
2913    
2914    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2915        character classes. PCRE was not treating the sequence [:...:] as a
2916        character class unless the ... were all letters. Perl, however, seems to
2917        allow any characters between [: and :], though of course it rejects as
2918        unknown any "names" that contain non-letters, because all the known class
2919        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2920        for example, whereas PCRE did not - it did not recognize a POSIX character
2921        class. This seemed a bit dangerous, so the code has been changed to be
2922        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2923        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2924        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2925        Perl does, and where it didn't before.
2926    
2927    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2928        Windows environments %n is disabled by default.
2929    
2930    
2931    Version 7.4 21-Sep-07
2932    ---------------------
2933    
2934    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2935        means that a class such as [\s] counted as "explicit reference to CR or
2936        LF". That isn't really right - the whole point of the change was to try to
2937        help when there was an actual mention of one of the two characters. So now
2938        the change happens only if \r or \n (or a literal CR or LF) character is
2939        encountered.
2940    
2941    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2942        of both had grown to the point where there were only 3 bits left.
2943        Fortunately, there was spare space in the data structure, and so I have
2944        moved the internal flags into a new 16-bit field to free up more option
2945        bits.
2946    
2947    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2948        but did not set the internal JCHANGED flag - either of these is enough to
2949        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2950        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2951        start sets both bits.
2952    
2953    4.  Added options (at build time, compile time, exec time) to change \R from
2954        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2955    
2956    5.  doc/pcresyntax.html was missing from the distribution.
2957    
2958    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2959        compatibility, even though it is no longer used.
2960    
2961    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2962        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2963        windows.h file is present (where different names are used). [This was
2964        reversed later after testing - see 16 below.]
2965    
2966    8.  Changed all #include <config.h> to #include "config.h". There were also
2967        some further <pcre.h> cases that I changed to "pcre.h".
2968    
2969    9.  When pcregrep was used with the --colour option, it missed the line ending
2970        sequence off the lines that it output.
2971    
2972    10. It was pointed out to me that arrays of string pointers cause lots of
2973        relocations when a shared library is dynamically loaded. A technique of
2974        using a single long string with a table of offsets can drastically reduce
2975        these. I have refactored PCRE in four places to do this. The result is
2976        dramatic:
2977    
2978          Originally:                          290
2979          After changing UCP table:            187
2980          After changing error message table:   43
2981          After changing table of "verbs"       36
2982          After changing table of Posix names   22
2983    
2984        Thanks to the folks working on Gregex for glib for this insight.
2985    
2986    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2987        unicode-properties was also set.
2988    
2989    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2990    
2991    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2992        checked only for CRLF.
2993    
2994    14. Added casts to pcretest.c to avoid compiler warnings.
2995    
2996    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2997    
2998    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2999        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3000        entirely. This removes changes made in 7 above.
3001    
3002    17. The CMake files have been updated, and there is now more information about
3003        building with CMake in the NON-UNIX-USE document.
3004    
3005    
3006    Version 7.3 28-Aug-07
3007    ---------------------
3008    
3009     1. In the rejigging of the build system that eventually resulted in 7.1, the
3010        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3011        brackets there is not right, since it causes compilers to look for an
3012        installed pcre.h, not the version that is in the source that is being
3013        compiled (which of course may be different). I have changed it back to:
3014    
3015          #include "pcre.h"
3016    
3017        I have a vague recollection that the change was concerned with compiling in
3018        different directories, but in the new build system, that is taken care of
3019        by the VPATH setting the Makefile.
3020    
3021     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3022        when the subject happened to end in the byte 0x85 (e.g. if the last
3023        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3024        characters but of course it shouldn't be taken as a newline when it is part
3025        of another character. The bug was that, for an unlimited repeat of . in
3026        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3027        characters when looking for a newline.
3028    
3029     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3030    
3031     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3032        in debug output.
3033    
3034     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3035        long printing in the pcrecpp unittest when running under MinGW.
3036    
3037     6. ESC_K was left out of the EBCDIC table.
3038    
3039     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3040        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3041        limit also applies to "virtual nesting" when a pattern is recursive, and in
3042        this case 1000 isn't so big. I have been able to remove this limit at the
3043        expense of backing off one optimization in certain circumstances. Normally,
3044        when pcre_exec() would call its internal match() function recursively and
3045        immediately return the result unconditionally, it uses a "tail recursion"
3046        feature to save stack. However, when a subpattern that can match an empty
3047        string has an unlimited repetition quantifier, it no longer makes this
3048        optimization. That gives it a stack frame in which to save the data for
3049        checking that an empty string has been matched. Previously this was taken
3050        from the 1000-entry workspace that had been reserved. So now there is no
3051        explicit limit, but more stack is used.
3052    
3053     8. Applied Daniel's patches to solve problems with the import/export magic
3054        syntax that is required for Windows, and which was going wrong for the
3055        pcreposix and pcrecpp parts of the library. These were overlooked when this
3056        problem was solved for the main library.
3057    
3058     9. There were some crude static tests to avoid integer overflow when computing
3059        the size of patterns that contain repeated groups with explicit upper
3060        limits. As the maximum quantifier is 65535, the maximum group length was
3061        set at 30,000 so that the product of these two numbers did not overflow a
3062        32-bit integer. However, it turns out that people want to use groups that
3063        are longer than 30,000 bytes (though not repeat them that many times).
3064        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3065        made it possible to implement the integer overflow checks in a much more
3066        dynamic way, which I have now done. The artificial limitation on group
3067        length has been removed - we now have only the limit on the total length of
3068        the compiled pattern, which depends on the LINK_SIZE setting.
3069    
3070    10. Fixed a bug in the documentation for get/copy named substring when
3071        duplicate names are permitted. If none of the named substrings are set, the
3072        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3073        empty string.
3074    
3075    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3076        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3077        because the ] is interpreted as the first data character and the
3078        terminating ] is not found. PCRE has been made compatible with Perl in this
3079        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3080        cause memory overwriting.
3081    
3082    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3083        string has been matched (to stop an infinite loop). It was not recognizing
3084        a conditional subpattern that could match an empty string if that
3085        subpattern was within another subpattern. For example, it looped when
3086        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3087        condition was not nested. This bug has been fixed.
3088    
3089    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3090        past the start of the subject in the presence of bytes with the top bit
3091        set, for example "\x8aBCD".
3092    
3093    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3094        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3095    
3096    14. Optimized (?!) to (*FAIL).
3097    
3098    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3099        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3100        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3101        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3102        does: it's just the validity check that is more restrictive.
3103    
3104    16. Inserted checks for integer overflows during escape sequence (backslash)
3105        processing, and also fixed erroneous offset values for syntax errors during
3106        backslash processing.
3107    
3108    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3109        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3110    
3111    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3112        caused an overrun.
3113    
3114    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3115        something other than just ASCII characters) inside a group that had an
3116        unlimited repeat caused a loop at compile time (while checking to see
3117        whether the group could match an empty string).
3118    
3119    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3120        [\P{Any}] did so. (Error in the code for printing property names.)
3121    
3122    21. An orphan \E inside a character class could cause a crash.
3123    
3124    22. A repeated capturing bracket such as (A)? could cause a wild memory
3125        reference during compilation.
3126    
3127    23. There are several functions in pcre_compile() that scan along a compiled
3128        expression for various reasons (e.g. to see if it's fixed length for look
3129        behind). There were bugs in these functions when a repeated \p or \P was
3130        present in the pattern. These operators have additional parameters compared
3131        with \d, etc, and these were not being taken into account when moving along
3132        the compiled data. Specifically:
3133    
3134        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3135            length.
3136    
3137        (b) An item such as \pL+ within a repeated group could cause crashes or
3138            loops.
3139    
3140        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3141            "reference to non-existent subpattern" error.
3142    
3143        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3144    
3145    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3146        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3147    
3148    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3149    
3150    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3151        character were causing crashes (broken optimization).
3152    
3153    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3154        \p or \P) caused a compile-time loop.
3155    
3156    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3157        break. For example, the unstudied pattern [\r\n]A does not match the string
3158        "\r\nA" because change 7.0/46 below moves the current point on by two
3159        characters after failing to match at the start. However, the pattern \nA
3160        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3161        the same is true. There doesn't seem any very clean way out of this, but
3162        what I have chosen to do makes the common cases work: PCRE now takes note
3163        of whether there can be an explicit match for \r or \n anywhere in the
3164        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3165        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3166        pattern has explicit CR or LF references.
3167    
3168    29. Added (*CR) etc for changing newline setting at start of pattern.
3169    
3170    
3171    Version 7.2 19-Jun-07
3172  ---------------------  ---------------------
3173    
3174   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 21  Version 7.2 01-May-07 Line 3188  Version 7.2 01-May-07
3188      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3189      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
3190      in all cases.      in all cases.
3191    
3192   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3193    
3194      (a) (?-n) (where n is a string of digits) is a relative subroutine or      (a) (?-n) (where n is a string of digits) is a relative subroutine or
3195          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3196    
3197      (b) (?+n) is also a relative subroutine call; it refers to the nth next      (b) (?+n) is also a relative subroutine call; it refers to the nth next
3198          to be opened parentheses.          to be opened parentheses.
3199    
3200      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3201          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3202    
3203        (d) \K resets the start of the current match so that everything before
3204            is not part of it.
3205    
3206        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3207    
3208        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3209            reference syntax.
3210    
3211        (g) (?| introduces a group in which the numbering of parentheses in each
3212            alternative starts with the same number.
3213    
3214        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3215    
3216     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3217        PCRE_INFO_JCHANGED.
3218    
3219     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3220        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3221        for detecting groups that can match an empty string.
3222    
3223     9. A pattern with a very large number of alternatives (more than several
3224        hundred) was running out of internal workspace during the pre-compile
3225        phase, where pcre_compile() figures out how much memory will be needed. A
3226        bit of new cunning has reduced the workspace needed for groups with
3227        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3228        workspace instead of running out of the 4096 that are available.
3229    
3230    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3231    
3232    11. Applied patch from Google to remove an optimization that didn't quite work.
3233        The report of the bug said:
3234    
3235          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3236          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3237          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3238    
3239    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3240        it matched the wrong number of bytes.
3241    
3242    
3243  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
3244  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1616

  ViewVC Help
Powered by ViewVC 1.1.5