/[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 1624 by zherczeg, Fri Feb 5 13:47:43 2016 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    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
31        space or a #-type comment that was followed by (?-x), which turns off
32        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
33        pcre_compile() assumed that (?-x) applied to the whole pattern and
34        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
35    
36    8.  An call of pcre_copy_named_substring() for a named substring whose number
37        was greater than the space in the ovector could cause a crash.
38    
39    9.  Yet another buffer overflow bug involved duplicate named groups with a
40        group that reset capture numbers (compare 8.38/7 below). Once again, I have
41        just allowed for more memory, even if not needed. (A proper fix is
42        implemented in PCRE2, but it involves a lot of refactoring.)
43    
44    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
45        start of the match to be earlier than the end.
46    
47    11. JIT code generator for assertion matching is refactored to a separate
48        function.
49    
50    12. Assertion code generator in JIT has been optimized.
51    
52    
53    Version 8.38 23-November-2015
54    -----------------------------
55    
56    1.  If a group that contained a recursive back reference also contained a
57        forward reference subroutine call followed by a non-forward-reference
58        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
59        compile correct code, leading to undefined behaviour or an internally
60        detected error. This bug was discovered by the LLVM fuzzer.
61    
62    2.  Quantification of certain items (e.g. atomic back references) could cause
63        incorrect code to be compiled when recursive forward references were
64        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
65        This bug was discovered by the LLVM fuzzer.
66    
67    3.  A repeated conditional group whose condition was a reference by name caused
68        a buffer overflow if there was more than one group with the given name.
69        This bug was discovered by the LLVM fuzzer.
70    
71    4.  A recursive back reference by name within a group that had the same name as
72        another group caused a buffer overflow. For example:
73        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
74    
75    5.  A forward reference by name to a group whose number is the same as the
76        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
77        a buffer overflow at compile time. This bug was discovered by the LLVM
78        fuzzer.
79    
80    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
81        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
82    
83    7.  Another buffer overflow bug involved duplicate named groups with a
84        reference between their definition, with a group that reset capture
85        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
86        fixed by always allowing for more memory, even if not needed. (A proper fix
87        is implemented in PCRE2, but it involves more refactoring.)
88    
89    8.  There was no check for integer overflow in subroutine calls such as (?123).
90    
91    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
92        being treated as a literal 'l' instead of causing an error.
93    
94    10. There was a buffer overflow if pcre_exec() was called with an ovector of
95        size 1. This bug was found by american fuzzy lop.
96    
97    11. If a non-capturing group containing a conditional group that could match
98        an empty string was repeated, it was not identified as matching an empty
99        string itself. For example: /^(?:(?(1)x|)+)+$()/.
100    
101    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
102        \a and \e in test subject lines.
103    
104    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
105        instead of the EBCDIC value.
106    
107    14. The handling of \c in an EBCDIC environment has been revised so that it is
108        now compatible with the specification in Perl's perlebcdic page.
109    
110    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
111        ASCII/Unicode. This has now been added to the list of characters that are
112        recognized as white space in EBCDIC.
113    
114    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
115        error (correctly) when used outside a class, but did not give an error
116        within a class.
117    
118    17. \h within a class was incorrectly compiled in EBCDIC environments.
119    
120    18. A pattern with an unmatched closing parenthesis that contained a backward
121        assertion which itself contained a forward reference caused buffer
122        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
123    
124    19. JIT should return with error when the compiled pattern requires more stack
125        space than the maximum.
126    
127    20. A possessively repeated conditional group that could match an empty string,
128        for example, /(?(R))*+/, was incorrectly compiled.
129    
130    21. Fix infinite recursion in the JIT compiler when certain patterns such as
131        /(?:|a|){100}x/ are analysed.
132    
133    22. Some patterns with character classes involving [: and \\ were incorrectly
134        compiled and could cause reading from uninitialized memory or an incorrect
135        error diagnosis.
136    
137    23. Pathological patterns containing many nested occurrences of [: caused
138        pcre_compile() to run for a very long time.
139    
140    24. A conditional group with only one branch has an implicit empty alternative
141        branch and must therefore be treated as potentially matching an empty
142        string.
143    
144    25. If (?R was followed by - or + incorrect behaviour happened instead of a
145        diagnostic.
146    
147    26. Arrange to give up on finding the minimum matching length for overly
148        complex patterns.
149    
150    27. Similar to (4) above: in a pattern with duplicated named groups and an
151        occurrence of (?| it is possible for an apparently non-recursive back
152        reference to become recursive if a later named group with the relevant
153        number is encountered. This could lead to a buffer overflow. Wen Guanxing
154        from Venustech ADLAB discovered this bug.
155    
156    28. If pcregrep was given the -q option with -c or -l, or when handling a
157        binary file, it incorrectly wrote output to stdout.
158    
159    29. The JIT compiler did not restore the control verb head in case of *THEN
160        control verbs. This issue was found by Karl Skomski with a custom LLVM
161        fuzzer.
162    
163    30. Error messages for syntax errors following \g and \k were giving inaccurate
164        offsets in the pattern.
165    
166    31. Added a check for integer overflow in conditions (?(<digits>) and
167        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
168        fuzzer.
169    
170    32. Handling recursive references such as (?2) when the reference is to a group
171        later in the pattern uses code that is very hacked about and error-prone.
172        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
173        give an internal error if it is obvious that compiling has gone wrong.
174    
175    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
176        This issue was found by Karl Skomski with a custom LLVM fuzzer.
177    
178    34. The JIT compiler should restore the control chain for empty possessive
179        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
180    
181    35. Match limit check added to JIT recursion. This issue was found by Karl
182        Skomski with a custom LLVM fuzzer.
183    
184    36. Yet another case similar to 27 above has been circumvented by an
185        unconditional allocation of extra memory. This issue is fixed "properly" in
186        PCRE2 by refactoring the way references are handled. Wen Guanxing
187        from Venustech ADLAB discovered this bug.
188    
189    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
190        with a custom LLVM fuzzer.
191    
192    38. Fixed a corner case of range optimization in JIT.
193    
194    39. An incorrect error "overran compiling workspace" was given if there were
195        exactly enough group forward references such that the last one extended
196        into the workspace safety margin. The next one would have expanded the
197        workspace. The test for overflow was not including the safety margin.
198    
199    40. A match limit issue is fixed in JIT which was found by Karl Skomski
200        with a custom LLVM fuzzer.
201    
202    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
203        work under Windows. (Why has it taken so long for anyone to notice?)
204    
205    42. In a character class such as [\W\p{Any}] where both a negative-type escape
206        ("not a word character") and a property escape were present, the property
207        escape was being ignored.
208    
209    43. Fix crash caused by very long (*MARK) or (*THEN) names.
210    
211    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
212        by a single ASCII character in a class item, was incorrectly compiled in
213        UCP mode. The POSIX class got lost, but only if the single character
214        followed it.
215    
216    45. [:punct:] in UCP mode was matching some characters in the range 128-255
217        that should not have been matched.
218    
219    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
220        class, all characters with code points greater than 255 are in the class.
221        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
222        such as \w are turned into Unicode properties), wide characters were not
223        correctly handled, and could fail to match.
224    
225    
226    Version 8.37 28-April-2015
227    --------------------------
228    
229    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
230        for those parentheses to be closed with whatever has been captured so far.
231        However, it was failing to mark any other groups between the hightest
232        capture so far and the currrent group as "unset". Thus, the ovector for
233        those groups contained whatever was previously there. An example is the
234        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
235    
236    2.  If an assertion condition was quantified with a minimum of zero (an odd
237        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
238    
239    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
240        unrecognized modifier, a crash could occur.
241    
242    4.  An attempt to do global matching in pcretest with a zero-length ovector
243        caused a crash.
244    
245    5.  Fixed a memory leak during matching that could occur for a subpattern
246        subroutine call (recursive or otherwise) if the number of captured groups
247        that had to be saved was greater than ten.
248    
249    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
250        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
251        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
252    
253    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
254        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
255    
256    8. When a pattern is compiled, it remembers the highest back reference so that
257       when matching, if the ovector is too small, extra memory can be obtained to
258       use instead. A conditional subpattern whose condition is a check on a
259       capture having happened, such as, for example in the pattern
260       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
261       setting the highest backreference number. This mattered only if pcre_exec()
262       was called with an ovector that was too small to hold the capture, and there
263       was no other kind of back reference (a situation which is probably quite
264       rare). The effect of the bug was that the condition was always treated as
265       FALSE when the capture could not be consulted, leading to a incorrect
266       behaviour by pcre_exec(). This bug has been fixed.
267    
268    9. A reference to a duplicated named group (either a back reference or a test
269       for being set in a conditional) that occurred in a part of the pattern where
270       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
271       to be incorrectly calculated, leading to overwriting.
272    
273    10. A mutually recursive set of back references such as (\2)(\1) caused a
274        segfault at study time (while trying to find the minimum matching length).
275        The infinite loop is now broken (with the minimum length unset, that is,
276        zero).
277    
278    11. If an assertion that was used as a condition was quantified with a minimum
279        of zero, matching went wrong. In particular, if the whole group had
280        unlimited repetition and could match an empty string, a segfault was
281        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
282        assertions to be quantified, but not if they are being used as conditions,
283        so the above pattern is faulted by Perl. PCRE has now been changed so that
284        it also rejects such patterns.
285    
286    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
287        failed to allow the zero-repeat case if pcre2_exec() was called with an
288        ovector too small to capture the group.
289    
290    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
291        Red Hat Product Security:
292    
293        (a) A crash if /K and /F were both set with the option to save the compiled
294        pattern.
295    
296        (b) Another crash if the option to print captured substrings in a callout
297        was combined with setting a null ovector, for example \O\C+ as a subject
298        string.
299    
300    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
301        forward reference repeated a large (but limited) number of times within a
302        repeated outer group that has a zero minimum quantifier, caused incorrect
303        code to be compiled, leading to the error "internal error:
304        previously-checked referenced subpattern not found" when an incorrect
305        memory address was read. This bug was reported as "heap overflow",
306        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
307        CVE-2015-2325.
308    
309    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
310        call within a group that also contained a recursive back reference caused
311        incorrect code to be compiled. This bug was reported as "heap overflow",
312        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
313        number CVE-2015-2326.
314    
315    24. Computing the size of the JIT read-only data in advance has been a source
316        of various issues, and new ones are still appear unfortunately. To fix
317        existing and future issues, size computation is eliminated from the code,
318        and replaced by on-demand memory allocation.
319    
320    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
321        adjacent to the end of the range, and the range contained characters with
322        more than one other case, caused incorrect behaviour when compiled in UTF
323        mode. In that example, the range a-j was left out of the class.
324    
325    26. Fix JIT compilation of conditional blocks, which assertion
326        is converted to (*FAIL). E.g: /(?(?!))/.
327    
328    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
329        discovered by the LLVM fuzzer.
330    
331    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
332        when this assertion was used as a condition, for example (?(?!)a|b). In
333        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
334        error about an unsupported item.
335    
336    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
337        possessification code could take exponential time to complete. A recursion
338        depth limit of 1000 has been imposed to limit the resources used by this
339        optimization.
340    
341    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
342        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
343        because \S ensures they are all in the class. The code for doing this was
344        interacting badly with the code for computing the amount of space needed to
345        compile the pattern, leading to a buffer overflow. This bug was discovered
346        by the LLVM fuzzer.
347    
348    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
349        other kinds of group caused stack overflow at compile time. This bug was
350        discovered by the LLVM fuzzer.
351    
352    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
353        between a subroutine call and its quantifier was incorrectly compiled,
354        leading to buffer overflow or other errors. This bug was discovered by the
355        LLVM fuzzer.
356    
357    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
358        assertion after (?(. The code was failing to check the character after
359        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
360        was discovered by the LLVM fuzzer.
361    
362    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
363        a fixed maximum following a group that contains a subroutine reference was
364        incorrectly compiled and could trigger buffer overflow. This bug was
365        discovered by the LLVM fuzzer.
366    
367    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
368        caused a stack overflow instead of the diagnosis of a non-fixed length
369        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
370    
371    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
372        (e.g. /(?<=\Ka)/) could make pcregrep loop.
373    
374    37. There was a similar problem to 36 in pcretest for global matches.
375    
376    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
377        and a subsequent item in the pattern caused a non-match, backtracking over
378        the repeated \X did not stop, but carried on past the start of the subject,
379        causing reference to random memory and/or a segfault. There were also some
380        other cases where backtracking after \C could crash. This set of bugs was
381        discovered by the LLVM fuzzer.
382    
383    39. The function for finding the minimum length of a matching string could take
384        a very long time if mutual recursion was present many times in a pattern,
385        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
386        method has been implemented. This infelicity was discovered by the LLVM
387        fuzzer.
388    
389    40. Static linking against the PCRE library using the pkg-config module was
390        failing on missing pthread symbols.
391    
392    
393    Version 8.36 26-September-2014
394    ------------------------------
395    
396    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
397        -Wmissing-field-initializers and -Wunused-parameter.
398    
399    2.  The tests for quantifiers being too big (greater than 65535) were being
400        applied after reading the number, and stupidly assuming that integer
401        overflow would give a negative number. The tests are now applied as the
402        numbers are read.
403    
404    3.  Tidy code in pcre_exec.c where two branches that used to be different are
405        now the same.
406    
407    4.  The JIT compiler did not generate match limit checks for certain
408        bracketed expressions with quantifiers. This may lead to exponential
409        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
410        issue should be resolved now.
411    
412    5.  Fixed an issue, which occures when nested alternatives are optimized
413        with table jumps.
414    
415    6.  Inserted two casts and changed some ints to size_t in the light of some
416        reported 64-bit compiler warnings (Bugzilla 1477).
417    
418    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
419        an empty string, which sometimes were behaving incorrectly in the
420        interpreter (though correctly in the JIT matcher). This pcretest input is
421        an example:
422    
423          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
424          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
425    
426        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
427        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
428        for an empty string was breaking the inner loop and carrying on at a lower
429        level, when possessive repeated groups should always return to a higher
430        level as they have no backtrack points in them. The empty string test now
431        occurs at the outer level.
432    
433    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
434        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
435    
436    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
437        doesn't).
438    
439    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
440        to hold a pattern to be incorrectly computed (too small) when there were
441        named back references to duplicated names. This could cause "internal
442        error: code overflow" or "double free or corruption" or other memory
443        handling errors.
444    
445    11. When named subpatterns had the same prefixes, back references could be
446        confused. For example, in this pattern:
447    
448          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
449    
450        the reference to 'Name' was incorrectly treated as a reference to a
451        duplicate name.
452    
453    12. A pattern such as /^s?c/mi8 where the optional character has more than
454        one "other case" was incorrectly compiled such that it would only try to
455        match starting at "c".
456    
457    13. When a pattern starting with \s was studied, VT was not included in the
458        list of possible starting characters; this should have been part of the
459        8.34/18 patch.
460    
461    14. If a character class started [\Qx]... where x is any character, the class
462        was incorrectly terminated at the ].
463    
464    15. If a pattern that started with a caseless match for a character with more
465        than one "other case" was studied, PCRE did not set up the starting code
466        unit bit map for the list of possible characters. Now it does. This is an
467        optimization improvement, not a bug fix.
468    
469    16. The Unicode data tables have been updated to Unicode 7.0.0.
470    
471    17. Fixed a number of memory leaks in pcregrep.
472    
473    18. Avoid a compiler warning (from some compilers) for a function call with
474        a cast that removes "const" from an lvalue by using an intermediate
475        variable (to which the compiler does not object).
476    
477    19. Incorrect code was compiled if a group that contained an internal recursive
478        back reference was optional (had quantifier with a minimum of zero). This
479        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
480        caused segmentation faults because of stack overflows at compile time.
481    
482    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
483        group that is quantified with an indefinite repeat, caused a compile-time
484        loop which used up all the system stack and provoked a segmentation fault.
485        This was not the same bug as 19 above.
486    
487    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
488        Patch by Mike Frysinger.
489    
490    
491    Version 8.35 04-April-2014
492    --------------------------
493    
494    1.  A new flag is set, when property checks are present in an XCLASS.
495        When this flag is not set, PCRE can perform certain optimizations
496        such as studying these XCLASS-es.
497    
498    2.  The auto-possessification of character sets were improved: a normal
499        and an extended character set can be compared now. Furthermore
500        the JIT compiler optimizes more character set checks.
501    
502    3.  Got rid of some compiler warnings for potentially uninitialized variables
503        that show up only when compiled with -O2.
504    
505    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
506        of a match later then the end of the match. The pcretest program was not
507        handling the case sensibly - it was outputting from the start to the next
508        binary zero. It now reports this situation in a message, and outputs the
509        text from the end to the start.
510    
511    5.  Fast forward search is improved in JIT. Instead of the first three
512        characters, any three characters with fixed position can be searched.
513        Search order: first, last, middle.
514    
515    6.  Improve character range checks in JIT. Characters are read by an inprecise
516        function now, which returns with an unknown value if the character code is
517        above a certain threshold (e.g: 256). The only limitation is that the value
518        must be bigger than the threshold as well. This function is useful when
519        the characters above the threshold are handled in the same way.
520    
521    7.  The macros whose names start with RAWUCHAR are placeholders for a future
522        mode in which only the bottom 21 bits of 32-bit data items are used. To
523        make this more memorable for those maintaining the code, the names have
524        been changed to start with UCHAR21, and an extensive comment has been added
525        to their definition.
526    
527    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
528        to the export list in Makefile.am (they were accidentally omitted from the
529        8.34 tarball).
530    
531    9.  The informational output from pcretest used the phrase "starting byte set"
532        which is inappropriate for the 16-bit and 32-bit libraries. As the output
533        for "first char" and "need char" really means "non-UTF-char", I've changed
534        "byte" to "char", and slightly reworded the output. The documentation about
535        these values has also been (I hope) clarified.
536    
537    10. Another JIT related optimization: use table jumps for selecting the correct
538        backtracking path, when more than four alternatives are present inside a
539        bracket.
540    
541    11. Empty match is not possible, when the minimum length is greater than zero,
542        and there is no \K in the pattern. JIT should avoid empty match checks in
543        such cases.
544    
545    12. In a caseless character class with UCP support, when a character with more
546        than one alternative case was not the first character of a range, not all
547        the alternative cases were added to the class. For example, s and \x{17f}
548        are both alternative cases for S: the class [RST] was handled correctly,
549        but [R-T] was not.
550    
551    13. The configure.ac file always checked for pthread support when JIT was
552        enabled. This is not used in Windows, so I have put this test inside a
553        check for the presence of windows.h (which was already tested for).
554    
555    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
556        The algorithm provides a way to skip certain starting offsets, and usually
557        faster than linear prefix searches.
558    
559    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
560        as for 'fr_FR' and 'french'. For some reason, however, it then used the
561        Windows-specific input and output files, which have 'french' screwed in.
562        So this could never have worked. One of the problems with locales is that
563        they aren't always the same. I have now updated RunTest so that it checks
564        the output of the locale test (test 3) against three different output
565        files, and it allows the test to pass if any one of them matches. With luck
566        this should make the test pass on some versions of Solaris where it was
567        failing. Because of the uncertainty, the script did not used to stop if
568        test 3 failed; it now does. If further versions of a French locale ever
569        come to light, they can now easily be added.
570    
571    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
572        there was a message during ./configure, but it did not stop. This now
573        provokes an error. The invalid example in README has been corrected.
574        If a value less than the minimum is given, the minimum value has always
575        been used, but now a warning is given.
576    
577    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
578        was a bug in the test system, which is now fixed. Also, the list of various
579        configurations that are tested for each release did not have one with both
580        16/32 bits and --enable-bar-anycrlf. It now does.
581    
582    18. pcretest was missing "-C bsr" for displaying the \R default setting.
583    
584    19. Little endian PowerPC systems are supported now by the JIT compiler.
585    
586    20. The fast forward newline mechanism could enter to an infinite loop on
587        certain invalid UTF-8 input. Although we don't support these cases
588        this issue can be fixed by a performance optimization.
589    
590    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
591        not take account if existing stack usage. There is now a new global
592        variable called pcre_stack_guard that can be set to point to an external
593        function to check stack availability. It is called at the start of
594        processing every parenthesized group.
595    
596    22. A typo in the code meant that in ungreedy mode the max/min qualifier
597        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
598        not match "ab".
599    
600    23. When UTF was disabled, the JIT program reported some incorrect compile
601        errors. These messages are silenced now.
602    
603    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
604        compiler.
605    
606    25. Change all the temporary files used in RunGrepTest to be different to those
607        used by RunTest so that the tests can be run simultaneously, for example by
608        "make -j check".
609    
610    
611    Version 8.34 15-December-2013
612    -----------------------------
613    
614    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
615        executable memory. Patch inspired by Carsten Klein.
616    
617    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
618        this macro is never tested and has no effect, because the work to support
619        coverage involves only compiling and linking options and special targets in
620        the Makefile. The comment in config.h implied that defining the macro would
621        enable coverage support, which is totally false. There was also support for
622        setting this macro in the CMake files (my fault, I just copied it from
623        configure). SUPPORT_GCOV has now been removed.
624    
625    3.  Make a small performance improvement in strlen16() and strlen32() in
626        pcretest.
627    
628    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
629        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
630        these cases). There was also one in pcretest.c.
631    
632    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
633    
634    6.  In UTF mode, the code for checking whether a group could match an empty
635        string (which is used for indefinitely repeated groups to allow for
636        breaking an infinite loop) was broken when the group contained a repeated
637        negated single-character class with a character that occupied more than one
638        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
639        UTF-8 mode). The effect was undefined: the group might or might not be
640        deemed as matching an empty string, or the program might have crashed.
641    
642    7.  The code for checking whether a group could match an empty string was not
643        recognizing that \h, \H, \v, \V, and \R must match a character.
644    
645    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
646        an empty string. If it can, pcretest shows this in its information output.
647    
648    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
649        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
650        matched by pcre_exec() without using JIT:
651    
652        (a) If the rest of the pattern did not match after a maximal run of
653            grapheme clusters, the code for backing up to try with fewer of them
654            did not always back up over a full grapheme when characters that do not
655            have the modifier quality were involved, e.g. Hangul syllables.
656    
657        (b) If the match point in a subject started with modifier character, and
658            there was no match, the code could incorrectly back up beyond the match
659            point, and potentially beyond the first character in the subject,
660            leading to a segfault or an incorrect match result.
661    
662    10. A conditional group with an assertion condition could lead to PCRE
663        recording an incorrect first data item for a match if no other first data
664        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
665        first data item, and therefore matched "ca" after "c" instead of at the
666        start.
667    
668    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
669        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
670    
671    12. The source of pcregrep now includes z/OS-specific code so that it can be
672        compiled for z/OS as part of the special z/OS distribution.
673    
674    13. Added the -T and -TM options to pcretest.
675    
676    14. The code in pcre_compile.c for creating the table of named capturing groups
677        has been refactored. Instead of creating the table dynamically during the
678        actual compiling pass, the information is remembered during the pre-compile
679        pass (on the stack unless there are more than 20 named groups, in which
680        case malloc() is used) and the whole table is created before the actual
681        compile happens. This has simplified the code (it is now nearly 150 lines
682        shorter) and prepared the way for better handling of references to groups
683        with duplicate names.
684    
685    15. A back reference to a named subpattern when there is more than one of the
686        same name now checks them in the order in which they appear in the pattern.
687        The first one that is set is used for the reference. Previously only the
688        first one was inspected. This change makes PCRE more compatible with Perl.
689    
690    16. Unicode character properties were updated from Unicode 6.3.0.
691    
692    17. The compile-time code for auto-possessification has been refactored, based
693        on a patch by Zoltan Herczeg. It now happens after instead of during
694        compilation. The code is cleaner, and more cases are handled. The option
695        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
696        options in pcretest are provided to set it. It can also be set by
697        (*NO_AUTO_POSSESS) at the start of a pattern.
698    
699    18. The character VT has been added to the default ("C" locale) set of
700        characters that match \s and are generally treated as white space,
701        following this same change in Perl 5.18. There is now no difference between
702        "Perl space" and "POSIX space". Whether VT is treated as white space in
703        other locales depends on the locale.
704    
705    19. The code for checking named groups as conditions, either for being set or
706        for being recursed, has been refactored (this is related to 14 and 15
707        above). Processing unduplicated named groups should now be as fast at
708        numerical groups, and processing duplicated groups should be faster than
709        before.
710    
711    20. Two patches to the CMake build system, by Alexander Barkov:
712    
713          (1) Replace the "source" command by "." in CMakeLists.txt because
714              "source" is a bash-ism.
715    
716          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
717              without these the CMake build does not work on Solaris.
718    
719    21. Perl has changed its handling of \8 and \9. If there is no previously
720        encountered capturing group of those numbers, they are treated as the
721        literal characters 8 and 9 instead of a binary zero followed by the
722        literals. PCRE now does the same.
723    
724    22. Following Perl, added \o{} to specify codepoints in octal, making it
725        possible to specify values greater than 0777 and also making them
726        unambiguous.
727    
728    23. Perl now gives an error for missing closing braces after \x{... instead of
729        treating the string as literal. PCRE now does the same.
730    
731    24. RunTest used to grumble if an inappropriate test was selected explicitly,
732        but just skip it when running all tests. This make it awkward to run ranges
733        of tests when one of them was inappropriate. Now it just skips any
734        inappropriate tests, as it always did when running all tests.
735    
736    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
737        character types such as \d or \w, too many callouts were inserted, and the
738        data that they returned was rubbish.
739    
740    26. In UCP mode, \s was not matching two of the characters that Perl matches,
741        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
742        were matched by \h. The code has now been refactored so that the lists of
743        the horizontal and vertical whitespace characters used for \h and \v (which
744        are defined only in one place) are now also used for \s.
745    
746    27. Add JIT support for the 64 bit TileGX architecture.
747        Patch by Jiong Wang (Tilera Corporation).
748    
749    28. Possessive quantifiers for classes (both explicit and automatically
750        generated) now use special opcodes instead of wrapping in ONCE brackets.
751    
752    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
753        (because it's meaningless), this was not happening when PCRE_CASELESS was
754        set. Not wrong, but inefficient.
755    
756    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
757        properties for \w, \d, etc) is present in a test regex. Otherwise if the
758        test contains no characters greater than 255, Perl doesn't realise it
759        should be using Unicode semantics.
760    
761    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
762        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
763        does in Unicode mode.
764    
765    32. Added the "forbid" facility to pcretest so that putting tests into the
766        wrong test files can sometimes be quickly detected.
767    
768    33. There is now a limit (default 250) on the depth of nesting of parentheses.
769        This limit is imposed to control the amount of system stack used at compile
770        time. It can be changed at build time by --with-parens-nest-limit=xxx or
771        the equivalent in CMake.
772    
773    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
774        errors. Perl warns for these when in warning mode, but PCRE has no facility
775        for giving warnings.
776    
777    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
778        However, this was not working for (?!) because it is optimized to (*FAIL),
779        for which PCRE does not allow quantifiers. The optimization is now disabled
780        when a quantifier follows (?!). I can't see any use for this, but it makes
781        things uniform.
782    
783    36. Perl no longer allows group names to start with digits, so I have made this
784        change also in PCRE. It simplifies the code a bit.
785    
786    37. In extended mode, Perl ignores spaces before a + that indicates a
787        possessive quantifier. PCRE allowed a space before the quantifier, but not
788        before the possessive +. It now does.
789    
790    38. The use of \K (reset reported match start) within a repeated possessive
791        group such as (a\Kb)*+ was not working.
792    
793    40. Document that the same character tables must be used at compile time and
794        run time, and that the facility to pass tables to pcre_exec() and
795        pcre_dfa_exec() is for use only with saved/restored patterns.
796    
797    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
798        features for Builds with MSVC:
799    
800        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
801           builds). The .rc files can be used to set FileDescription and many other
802           attributes.
803    
804        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
805           This allows higher-level build scripts which want .pdb files to avoid
806           hard-coding the exact files needed."
807    
808    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
809        mean "start of word" and "end of word", respectively, as a transition aid.
810    
811    43. A minimizing repeat of a class containing codepoints greater than 255 in
812        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
813        compiled to use the heap for recursion.
814    
815    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
816        is configured.
817    
818    
819    Version 8.33 28-May-2013
820    ------------------------
821    
822    1.  Added 'U' to some constants that are compared to unsigned integers, to
823        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
824        variables that are added to signed variables, to ensure the result is
825        signed and can be negated.
826    
827    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
828        CMake config files.
829    
830    3.  Revise the creation of config.h.generic so that all boolean macros are
831        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
832        overriding via -D on the command line possible.
833    
834    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
835        to unsigned int is reported to make a quite noticeable speed difference in
836        a specific Windows environment. Testing on Linux did also appear to show
837        some benefit (and it is clearly not harmful). Also fixed the definition of
838        Xop which should be unsigned.
839    
840    5.  Related to (4), changing the definition of the intermediate variable cc
841        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
842        improvements.
843    
844    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
845        unnecessary spaces.
846    
847    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
848        and later.
849    
850    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
851    
852    9.  Optimizing fast_forward_start_bits in JIT.
853    
854    10. Adding support for callouts in JIT, and fixing some issues revealed
855        during this work. Namely:
856    
857        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
858    
859        (b) Minimum length was not checked before the matching is started.
860    
861    11. The value of capture_last that is passed to callouts was incorrect in some
862        cases when there was a capture on one path that was subsequently abandoned
863        after a backtrack. Also, the capture_last value is now reset after a
864        recursion, since all captures are also reset in this case.
865    
866    12. The interpreter no longer returns the "too many substrings" error in the
867        case when an overflowing capture is in a branch that is subsequently
868        abandoned after a backtrack.
869    
870    13. In the pathological case when an offset vector of size 2 is used, pcretest
871        now prints out the matched string after a yield of 0 or 1.
872    
873    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
874        Only supported by the JIT compiler at the moment.
875    
876    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
877    
878    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
879        offset of the starting point of the matching process, provided the offsets
880        vector is large enough.
881    
882    17. The \A escape now records a lookbehind value of 1, though its execution
883        does not actually inspect the previous character. This is to ensure that,
884        in partial multi-segment matching, at least one character from the old
885        segment is retained when a new segment is processed. Otherwise, if there
886        are no lookbehinds in the pattern, \A might match incorrectly at the start
887        of a new segment.
888    
889    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
890    
891    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
892        gives some modest performance improvement in 8-bit mode.
893    
894    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
895        be expressed in certain programming languages using Universal Character
896        Names.
897    
898    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
899        which points out that "non characters" are not "characters that may not
900        appear in Unicode strings" but rather "characters that are reserved for
901        internal use and have only local meaning".
902    
903    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
904        there was a conditional group that depended on an assertion, if the
905        assertion was false, the callout that immediately followed the alternation
906        in the condition was skipped when pcre_exec() was used for matching.
907    
908    23. Allow an explicit callout to be inserted before an assertion that is the
909        condition for a conditional group, for compatibility with automatic
910        callouts, which always insert a callout at this point.
911    
912    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
913        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
914    
915    25. (*PRUNE) is now supported by the JIT compiler.
916    
917    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
918    
919    27. Fix the case where there are two or more SKIPs with arguments that may be
920        ignored.
921    
922    28. (*SKIP) is now supported by the JIT compiler.
923    
924    29. (*THEN) is now supported by the JIT compiler.
925    
926    30. Update RunTest with additional test selector options.
927    
928    31. The way PCRE handles backtracking verbs has been changed in two ways.
929    
930        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
931        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
932        backtracking. In some cases this makes it more Perl-compatible, but Perl's
933        rather obscure rules do not always do the same thing.
934    
935        (2) Previously, backtracking verbs were confined within assertions. This is
936        no longer the case for positive assertions, except for (*ACCEPT). Again,
937        this sometimes improves Perl compatibility, and sometimes does not.
938    
939    32. A number of tests that were in test 2 because Perl did things differently
940        have been moved to test 1, because either Perl or PCRE has changed, and
941        these tests are now compatible.
942    
943    32. Backtracking control verbs are now handled in the same way in JIT and
944        interpreter.
945    
946    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
947        contained a forward subroutine reference caused a compile error.
948    
949    34. Auto-detect and optimize limited repetitions in JIT.
950    
951    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
952        blocking (*UTF) etc.
953    
954    36. In the interpreter, maximizing pattern repetitions for characters and
955        character types now use tail recursion, which reduces stack usage.
956    
957    37. The value of the max lookbehind was not correctly preserved if a compiled
958        and saved regex was reloaded on a host of different endianness.
959    
960    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
961        of the compiled pattern block, expand the flags field from 16 to 32 bits
962        because it was almost full.
963    
964    39. Try madvise first before posix_madvise.
965    
966    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
967        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
968        original change locked out all matches of empty strings. This has been
969        changed so that one match of an empty string per line is recognized.
970        Subsequent searches on the same line (for colouring or for --only-matching,
971        for example) do not recognize empty strings.
972    
973    41. Applied a user patch to fix a number of spelling mistakes in comments.
974    
975    42. Data lines longer than 65536 caused pcretest to crash.
976    
977    43. Clarified the data type for length and startoffset arguments for pcre_exec
978        and pcre_dfa_exec in the function-specific man pages, where they were
979        explicitly stated to be in bytes, never having been updated. I also added
980        some clarification to the pcreapi man page.
981    
982    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
983        a segmentation fault.
984    
985    
986    Version 8.32 30-November-2012
987    -----------------------------
988    
989    1.  Improved JIT compiler optimizations for first character search and single
990        character iterators.
991    
992    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
993        Patch by Daniel Richard G.
994    
995    3.  Single character iterator optimizations in the JIT compiler.
996    
997    4.  Improved JIT compiler optimizations for character ranges.
998    
999    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1000        Reported by Giuseppe D'Angelo.
1001    
1002    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1003        of a line, was being set incorrectly in cases where .* appeared inside
1004        atomic brackets at the start of a pattern, or where there was a subsequent
1005        *PRUNE or *SKIP.
1006    
1007    7.  Improved instruction cache flush for POWER/PowerPC.
1008        Patch by Daniel Richard G.
1009    
1010    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1011        grep:
1012    
1013        (a) There is now no limit to the number of patterns to be matched.
1014    
1015        (b) An error is given if a pattern is too long.
1016    
1017        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1018            are now supported.
1019    
1020        (d) --exclude-from and --include-from (multiple use) have been added.
1021    
1022        (e) Exclusions and inclusions now apply to all files and directories, not
1023            just to those obtained from scanning a directory recursively.
1024    
1025        (f) Multiple uses of -f and --file-list are now supported.
1026    
1027        (g) In a Windows environment, the default for -d has been changed from
1028            "read" (the GNU grep default) to "skip", because otherwise the presence
1029            of a directory in the file list provokes an error.
1030    
1031        (h) The documentation has been revised and clarified in places.
1032    
1033    9.  Improve the matching speed of capturing brackets.
1034    
1035    10. Changed the meaning of \X so that it now matches a Unicode extended
1036        grapheme cluster.
1037    
1038    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1039        out POSIX threads when JIT support is configured.
1040    
1041    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1042    
1043    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1044        structure could go wrong in environments where size_t is not the same size
1045        as int.
1046    
1047    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1048        set.
1049    
1050    15. The EBCDIC support had decayed; later updates to the code had included
1051        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1052        general tidy up of EBCDIC-related issues, and the documentation was also
1053        not quite right. There is now a test that can be run on ASCII systems to
1054        check some of the EBCDIC-related things (but is it not a full test).
1055    
1056    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1057        in a small tidy to the code.
1058    
1059    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1060    
1061    18. If the --only-matching (-o) option in pcregrep is specified multiple
1062        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1063        substrings matched by the 1st and 2nd capturing parentheses. A separating
1064        string can be specified by --om-separator (default empty).
1065    
1066    19. Improving the first n character searches.
1067    
1068    20. Turn case lists for horizontal and vertical white space into macros so that
1069        they are defined only once.
1070    
1071    21. This set of changes together give more compatible Unicode case-folding
1072        behaviour for characters that have more than one other case when UCP
1073        support is available.
1074    
1075        (a) The Unicode property table now has offsets into a new table of sets of
1076            three or more characters that are case-equivalent. The MultiStage2.py
1077            script that generates these tables (the pcre_ucd.c file) now scans
1078            CaseFolding.txt instead of UnicodeData.txt for character case
1079            information.
1080    
1081        (b) The code for adding characters or ranges of characters to a character
1082            class has been abstracted into a generalized function that also handles
1083            case-independence. In UTF-mode with UCP support, this uses the new data
1084            to handle characters with more than one other case.
1085    
1086        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1087            whose other case is greater than 256 are now correctly matched
1088            caselessly. Previously, the high codepoint matched the low one, but not
1089            vice versa.
1090    
1091        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1092            of the new class addition function, using character lists defined as
1093            macros alongside the case definitions of 20 above.
1094    
1095        (e) Caseless back references now work with characters that have more than
1096            one other case.
1097    
1098        (f) General caseless matching of characters with more than one other case
1099            is supported.
1100    
1101    22. Unicode character properties were updated from Unicode 6.2.0
1102    
1103    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1104    
1105    24. Add support for 32-bit character strings, and UTF-32
1106    
1107    25. Major JIT compiler update (code refactoring and bugfixing).
1108        Experimental Sparc 32 support is added.
1109    
1110    26. Applied a modified version of Daniel Richard G's patch to create
1111        pcre.h.generic and config.h.generic by "make" instead of in the
1112        PrepareRelease script.
1113    
1114    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1115        pcre_compile.c when checking for a zero character.
1116    
1117    28. Introducing a native interface for JIT. Through this interface, the compiled
1118        machine code can be directly executed. The purpose of this interface is to
1119        provide fast pattern matching, so several sanity checks are not performed.
1120        However, feature tests are still performed. The new interface provides
1121        1.4x speedup compared to the old one.
1122    
1123    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1124        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1125        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1126    
1127    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1128        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1129        the "old" RFC 2279). Instead, it ended up passing a negative length to
1130        pcre_exec().
1131    
1132    31. Add support for GCC's visibility feature to hide internal functions.
1133    
1134    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1135        "unknown -C option" after outputting 0 or 1.
1136    
1137    33. There is now support for generating a code coverage report for the test
1138        suite in environments where gcc is the compiler and lcov is installed. This
1139        is mainly for the benefit of the developers.
1140    
1141    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1142        unaddressable using valgrind annotations, allowing valgrind to detect
1143        invalid memory accesses. This is mainly for the benefit of the developers.
1144    
1145    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1146    
1147    26. Give configure error if --enable-cpp but no C++ compiler found.
1148    
1149    
1150    Version 8.31 06-July-2012
1151    -------------------------
1152    
1153    1.  Fixing a wrong JIT test case and some compiler warnings.
1154    
1155    2.  Removed a bashism from the RunTest script.
1156    
1157    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1158        to unsigned type, result still unsigned" that was given by an MS compiler
1159        on encountering the code "-sizeof(xxx)".
1160    
1161    4.  Partial matching support is added to the JIT compiler.
1162    
1163    5.  Fixed several bugs concerned with partial matching of items that consist
1164        of more than one character:
1165    
1166        (a) /^(..)\1/ did not partially match "aba" because checking references was
1167            done on an "all or nothing" basis. This also applied to repeated
1168            references.
1169    
1170        (b) \R did not give a hard partial match if \r was found at the end of the
1171            subject.
1172    
1173        (c) \X did not give a hard partial match after matching one or more
1174            characters at the end of the subject.
1175    
1176        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1177            a partial match for the string "\r".
1178    
1179        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1180            a partial match for a CR character at the end of the subject string.
1181    
1182    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1183        running pcretest, the text "(JIT)" added to the output whenever JIT is
1184        actually used to run the match.
1185    
1186    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1187        or /S+[+] with a digit between 1 and 7.
1188    
1189    8.  OP_NOT now supports any UTF character not just single-byte ones.
1190    
1191    9.  (*MARK) control verb is now supported by the JIT compiler.
1192    
1193    10. The command "./RunTest list" lists the available tests without actually
1194        running any of them. (Because I keep forgetting what they all are.)
1195    
1196    11. Add PCRE_INFO_MAXLOOKBEHIND.
1197    
1198    12. Applied a (slightly modified) user-supplied patch that improves performance
1199        when the heap is used for recursion (compiled with --disable-stack-for-
1200        recursion). Instead of malloc and free for each heap frame each time a
1201        logical recursion happens, frames are retained on a chain and re-used where
1202        possible. This sometimes gives as much as 30% improvement.
1203    
1204    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1205        call.
1206    
1207    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1208    
1209    15. It is now possible to link pcretest with libedit as an alternative to
1210        libreadline.
1211    
1212    16. (*COMMIT) control verb is now supported by the JIT compiler.
1213    
1214    17. The Unicode data tables have been updated to Unicode 6.1.0.
1215    
1216    18. Added --file-list option to pcregrep.
1217    
1218    19. Added binary file support to pcregrep, including the -a, --binary-files,
1219        -I, and --text options.
1220    
1221    20. The madvise function is renamed for posix_madvise for QNX compatibility
1222        reasons. Fixed by Giuseppe D'Angelo.
1223    
1224    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1225    
1226    22. Changed the option for creating symbolic links for 16-bit man pages from
1227        -s to -sf so that re-installing does not cause issues.
1228    
1229    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1230    
1231    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1232        matches in certain environments (the workspace was not being correctly
1233        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1234        some of the workspace data at the beginning of a restart.
1235    
1236    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1237        was not doing so when it should - probably a typo introduced by SVN 528
1238        (change 8.10/14).
1239    
1240    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1241        \w+ when the character tables indicated that \x{c4} was a word character.
1242        There were several related cases, all because the tests for doing a table
1243        lookup were testing for characters less than 127 instead of 255.
1244    
1245    27. If a pattern contains capturing parentheses that are not used in a match,
1246        their slots in the ovector are set to -1. For those that are higher than
1247        any matched groups, this happens at the end of processing. In the case when
1248        there were back references that the ovector was too small to contain
1249        (causing temporary malloc'd memory to be used during matching), and the
1250        highest capturing number was not used, memory off the end of the ovector
1251        was incorrectly being set to -1. (It was using the size of the temporary
1252        memory instead of the true size.)
1253    
1254    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1255        ovector size, it uses memory at the end of the block that it has got.
1256    
1257    29. Check for an overlong MARK name and give an error at compile time. The
1258        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1259    
1260    30. JIT compiler update.
1261    
1262    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1263        Rill for the patch.
1264    
1265    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1266    
1267    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1268    
1269    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1270        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1271        was enabled.
1272    
1273    35. Improve JIT code generation for greedy plus quantifier.
1274    
1275    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1276        1 to "aa" instead of to an empty string. The bug affected repeated groups
1277        that could potentially match an empty string.
1278    
1279    37. Optimizing single character iterators in JIT.
1280    
1281    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1282        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1283        codepoints that are too big for the mode are faulted, and in a UTF mode,
1284        disallowed codepoints are also faulted.
1285    
1286    39. If PCRE was compiled with UTF support, in three places in the DFA
1287        matcher there was code that should only have been obeyed in UTF mode, but
1288        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1289        processing when bytes with values greater than 127 were present. In 16-bit
1290        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1291        both cases the values are those that cannot be the first data item in a UTF
1292        character. The three items that might have provoked this were recursions,
1293        possessively repeated groups, and atomic groups.
1294    
1295    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1296        and pcregrep, because some OS require shared objects to be explicitly
1297        passed to ld, causing the link step to fail if they are not.
1298    
1299    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1300        mode, patterns that started with \h* or \R* might be incorrectly matched.
1301    
1302    
1303    Version 8.30 04-February-2012
1304    -----------------------------
1305    
1306    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1307        name is defined in ctype.h.
1308    
1309    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1310        only in quite long subpatterns.
1311    
1312    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1313        since it was replaced by pcre_fullinfo() in February 2000.
1314    
1315    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1316        match a (*MARK), and the match failed at the start of the subject, a
1317        reference to memory before the start of the subject could occur. This bug
1318        was introduced by fix 17 of release 8.21.
1319    
1320    5.  A reference to an unset group with zero minimum repetition was giving
1321        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1322        /(another)?(\1?)test/ matched against "hello world test". This bug was
1323        introduced in release 8.13.
1324    
1325    6.  Add support for 16-bit character strings (a large amount of work involving
1326        many changes and refactorings).
1327    
1328    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1329        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1330        from a file.
1331    
1332    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1333        rounding is not applied in this particular case).
1334    
1335    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1336        if they appear, or are escaped, in patterns.
1337    
1338    10. Get rid of a number of -Wunused-but-set-variable warnings.
1339    
1340    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1341        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1342        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1343        also returns the mark "x". This bug applied to capturing parentheses,
1344        non-capturing parentheses, and atomic parentheses. It also applied to some
1345        assertions.
1346    
1347    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1348        information out of configure.ac instead of relying on pcre.h.generic, which
1349        is not stored in the repository.
1350    
1351    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1352        -lreadline.
1353    
1354    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1355    
1356    15. Applied Graycode's patch to put the top-level frame on the stack rather
1357        than the heap when not using the stack for recursion. This gives a
1358        performance improvement in many cases when recursion is not deep.
1359    
1360    16. Experimental code added to "pcretest -C" to output the stack frame size.
1361    
1362    
1363    Version 8.21 12-Dec-2011
1364    ------------------------
1365    
1366    1.  Updating the JIT compiler.
1367    
1368    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1369        are added as well.
1370    
1371    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1372        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1373        calling _pcre_jit_exec. Some extra comments are added.
1374    
1375    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1376        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1377        was introduced by change 18 for 8.20.
1378    
1379    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1380        ECMA-262 standard.
1381    
1382    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1383        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1384        This bug was probably introduced by change 9 of 8.13.
1385    
1386    7.  While fixing 6 above, I noticed that a number of other items were being
1387        incorrectly rejected as "not fixed length". This arose partly because newer
1388        opcodes had not been added to the fixed-length checking code. I have (a)
1389        corrected the bug and added tests for these items, and (b) arranged for an
1390        error to occur if an unknown opcode is encountered while checking for fixed
1391        length instead of just assuming "not fixed length". The items that were
1392        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1393        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1394        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1395    
1396    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1397        being incorrectly compiled and would have given unpredicatble results.
1398    
1399    9.  A possessively repeated subpattern with minimum repeat count greater than
1400        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1401        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1402        the first (A) could occur when it should not.
1403    
1404    10. Add a cast and remove a redundant test from the code.
1405    
1406    11. JIT should use pcre_malloc/pcre_free for allocation.
1407    
1408    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1409        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1410        is anything other than /usr, -L is still shown).
1411    
1412    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1413    
1414    14. Perl does not support \N without a following name in a [] class; PCRE now
1415        also gives an error.
1416    
1417    15. If a forward reference was repeated with an upper limit of around 2000,
1418        it caused the error "internal error: overran compiling workspace". The
1419        maximum number of forward references (including repeats) was limited by the
1420        internal workspace, and dependent on the LINK_SIZE. The code has been
1421        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1422        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1423        of around 200,000 forward references. While doing this, I also speeded up
1424        the filling in of repeated forward references.
1425    
1426    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1427        incorrectly expecting the subject to contain another "a" after the start.
1428    
1429    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1430        in the match, the SKIP should be ignored. This was not happening; instead
1431        the SKIP was being treated as NOMATCH. For patterns such as
1432        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1433        tested.
1434    
1435    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1436        now much more compatible with Perl, in particular in cases where the result
1437        is a non-match for a non-anchored pattern. For example, if
1438        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1439        "m", where previously it did not return a name. A side effect of this
1440        change is that for partial matches, the last encountered mark name is
1441        returned, as for non matches. A number of tests that were previously not
1442        Perl-compatible have been moved into the Perl-compatible test files. The
1443        refactoring has had the pleasing side effect of removing one argument from
1444        the match() function, thus reducing its stack requirements.
1445    
1446    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1447        subsequent uses of /S (without +) incorrectly behaved like /S+.
1448    
1449    21. Retrieve executable code size support for the JIT compiler and fixing
1450        some warnings.
1451    
1452    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1453        not work when the shorter character appeared right at the end of the
1454        subject string.
1455    
1456    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1457        systems.
1458    
1459    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1460        output it when the /M option is used in pcretest.
1461    
1462    25. The CheckMan script was not being included in the distribution. Also, added
1463        an explicit "perl" to run Perl scripts from the PrepareRelease script
1464        because this is reportedly needed in Windows.
1465    
1466    26. If study data was being save in a file and studying had not found a set of
1467        "starts with" bytes for the pattern, the data written to the file (though
1468        never used) was taken from uninitialized memory and so caused valgrind to
1469        complain.
1470    
1471    27. Updated RunTest.bat as provided by Sheri Pierce.
1472    
1473    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1474    
1475    29. Computation of memory usage for the table of capturing group names was
1476        giving an unnecessarily large value.
1477    
1478    
1479    Version 8.20 21-Oct-2011
1480    ------------------------
1481    
1482    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1483        a POSIX class. After further experiments with Perl, which convinced me that
1484        Perl has bugs and confusions, a closing square bracket is no longer allowed
1485        in a POSIX name. This bug also affected patterns with classes that started
1486        with full stops.
1487    
1488    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1489        captured substring, but while checking the failing first alternative,
1490        substring 1 is temporarily captured. If the output vector supplied to
1491        pcre_exec() was not big enough for this capture, the yield of the function
1492        was still zero ("insufficient space for captured substrings"). This cannot
1493        be totally fixed without adding another stack variable, which seems a lot
1494        of expense for a edge case. However, I have improved the situation in cases
1495        such as /(a)(b)x|abc/ matched against "abc", where the return code
1496        indicates that fewer than the maximum number of slots in the ovector have
1497        been set.
1498    
1499    3.  Related to (2) above: when there are more back references in a pattern than
1500        slots in the output vector, pcre_exec() uses temporary memory during
1501        matching, and copies in the captures as far as possible afterwards. It was
1502        using the entire output vector, but this conflicts with the specification
1503        that only 2/3 is used for passing back captured substrings. Now it uses
1504        only the first 2/3, for compatibility. This is, of course, another edge
1505        case.
1506    
1507    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1508        main code base, and can be used by building with --enable-jit. When this is
1509        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1510        runtime --no-jit option is given.
1511    
1512    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1513        ovector, the return from the function was zero, implying that there were
1514        other matches that did not fit. The correct "exactly full" value is now
1515        returned.
1516    
1517    6.  If a subpattern that was called recursively or as a subroutine contained
1518        (*PRUNE) or any other control that caused it to give a non-standard return,
1519        invalid errors such as "Error -26 (nested recursion at the same subject
1520        position)" or even infinite loops could occur.
1521    
1522    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1523        computing the minimum length on reaching *ACCEPT, and so ended up with the
1524        wrong value of 1 rather than 0. Further investigation indicates that
1525        computing a minimum subject length in the presence of *ACCEPT is difficult
1526        (think back references, subroutine calls), and so I have changed the code
1527        so that no minimum is registered for a pattern that contains *ACCEPT.
1528    
1529    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1530        it was not handled as intended. [But see 16 below.]
1531    
1532    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1533        Sheri Pierce.
1534    
1535    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1536        the first byte in a match must be "a".
1537    
1538    11. Change 17 for 8.13 increased the recursion depth for patterns like
1539        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1540        pattern contains any instances of (*THEN). If it does not, the old
1541        optimizations are restored. It would be nice to do this on a per-group
1542        basis, but at the moment that is not feasible.
1543    
1544    12. In some environments, the output of pcretest -C is CRLF terminated. This
1545        broke RunTest's code that checks for the link size. A single white space
1546        character after the value is now allowed for.
1547    
1548    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1549        For "fr", it uses the Windows-specific input and output files.
1550    
1551    14. If (*THEN) appeared in a group that was called recursively or as a
1552        subroutine, it did not work as intended. [But see next item.]
1553    
1554    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1555        pattern fragments (but not containing any | characters). If A and B are
1556        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1557        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1558        D. In other words, Perl considers parentheses that do not contain any |
1559        characters to be part of a surrounding alternative, whereas PCRE was
1560        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1561        differently. PCRE now behaves in the same way as Perl, except in the case
1562        of subroutine/recursion calls such as (?1) which have in any case always
1563        been different (but PCRE had them first :-).
1564    
1565    16. Related to 15 above: Perl does not treat the | in a conditional group as
1566        creating alternatives. Such a group is treated in the same way as an
1567        ordinary group without any | characters when processing (*THEN). PCRE has
1568        been changed to match Perl's behaviour.
1569    
1570    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1571        RunGrepTest script failed.
1572    
1573    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1574        inevitable for groups that contain captures, but it can lead to a lot of
1575        stack use in large patterns. The old behaviour has been restored for atomic
1576        groups that do not contain any capturing parentheses.
1577    
1578    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1579        suppress the check for a minimum subject length at run time. (If it was
1580        given to pcre_exec() or pcre_dfa_exec() it did work.)
1581    
1582    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1583        fail to work when decoding hex characters in data strings in EBCDIC
1584        environments.
1585    
1586    21. It appears that in at least one Mac OS environment, the isxdigit() function
1587        is implemented as a macro that evaluates to its argument more than once,
1588        contravening the C 90 Standard (I haven't checked a later standard). There
1589        was an instance in pcretest which caused it to go wrong when processing
1590        \x{...} escapes in subject strings. The has been rewritten to avoid using
1591        things like p++ in the argument of isxdigit().
1592    
1593    
1594    Version 8.13 16-Aug-2011
1595    ------------------------
1596    
1597    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1598    
1599    2.  Two minor typos in pcre_internal.h have been fixed.
1600    
1601    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1602        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1603        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1604    
1605    4.  There were a number of related bugs in the code for matching backrefences
1606        caselessly in UTF-8 mode when codes for the characters concerned were
1607        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1608        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1609        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1610        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1611        2-byte code at the end of the subject (it thought there wasn't enough data
1612        left).
1613    
1614    5.  Comprehensive information about what went wrong is now returned by
1615        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1616        as the output vector has at least 2 elements. The offset of the start of
1617        the failing character and a reason code are placed in the vector.
1618    
1619    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1620        now returned is for the first byte of the failing character, instead of the
1621        last byte inspected. This is an incompatible change, but I hope it is small
1622        enough not to be a problem. It makes the returned offset consistent with
1623        pcre_exec() and pcre_dfa_exec().
1624    
1625    7.  pcretest now gives a text phrase as well as the error number when
1626        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1627        failure, the offset and reason code are output.
1628    
1629    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1630        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1631        back over a single character (\n). This seems wrong (because it treated the
1632        two characters as a single entity when going forwards), conflicts with the
1633        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1634        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1635        has been changed.
1636    
1637    9.  Some internal refactoring has changed the processing so that the handling
1638        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1639        time (the PCRE_DOTALL option was changed this way some time ago: version
1640        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1641        which was always a bit of a fudge. It also means that there is one less
1642        argument for the match() function, which reduces its stack requirements
1643        slightly. This change also fixes an incompatibility with Perl: the pattern
1644        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1645    
1646    10. More internal refactoring has drastically reduced the number of recursive
1647        calls to match() for possessively repeated groups such as (abc)++ when
1648        using pcre_exec().
1649    
1650    11. While implementing 10, a number of bugs in the handling of groups were
1651        discovered and fixed:
1652    
1653        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1654        (a|)*(?1) gave a compile-time internal error.
1655        ((a|)+)+  did not notice that the outer group could match an empty string.
1656        (^a|^)+   was not marked as anchored.
1657        (.*a|.*)+ was not marked as matching at start or after a newline.
1658    
1659    12. Yet more internal refactoring has removed another argument from the match()
1660        function. Special calls to this function are now indicated by setting a
1661        value in a variable in the "match data" data block.
1662    
1663    13. Be more explicit in pcre_study() instead of relying on "default" for
1664        opcodes that mean there is no starting character; this means that when new
1665        ones are added and accidentally left out of pcre_study(), testing should
1666        pick them up.
1667    
1668    14. The -s option of pcretest has been documented for ages as being an old
1669        synonym of -m (show memory usage). I have changed it to mean "force study
1670        for every regex", that is, assume /S for every regex. This is similar to -i
1671        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1672        using it. It makes it easier to run collections of tests with and without
1673        study enabled, and thereby test pcre_study() more easily. All the standard
1674        tests are now run with and without -s (but some patterns can be marked as
1675        "never study" - see 20 below).
1676    
1677    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1678        restoration of the capturing data to the outer values was not happening
1679        correctly.
1680    
1681    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1682        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1683        pattern had matched an empty string, and so incorrectly returned a no
1684        match.
1685    
1686    17. There was optimizing code for the last branch of non-capturing parentheses,
1687        and also for the obeyed branch of a conditional subexpression, which used
1688        tail recursion to cut down on stack usage. Unfortunately, now that there is
1689        the possibility of (*THEN) occurring in these branches, tail recursion is
1690        no longer possible because the return has to be checked for (*THEN). These
1691        two optimizations have therefore been removed. [But see 8.20/11 above.]
1692    
1693    18. If a pattern containing \R was studied, it was assumed that \R always
1694        matched two bytes, thus causing the minimum subject length to be
1695        incorrectly computed because \R can also match just one byte.
1696    
1697    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1698        was incorrectly computed.
1699    
1700    20. If /S is present twice on a test pattern in pcretest input, it now
1701        *disables* studying, thereby overriding the use of -s on the command line
1702        (see 14 above). This is necessary for one or two tests to keep the output
1703        identical in both cases.
1704    
1705    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1706        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1707    
1708    22. When an atomic group that contained a capturing parenthesis was
1709        successfully matched, but the branch in which it appeared failed, the
1710        capturing was not being forgotten if a higher numbered group was later
1711        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1712        group 1 to "a", when in fact it should be unset. This applied to multi-
1713        branched capturing and non-capturing groups, repeated or not, and also to
1714        positive assertions (capturing in negative assertions does not happen
1715        in PCRE) and also to nested atomic groups.
1716    
1717    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1718        subject after a captured substring, to make it easier to tell which of a
1719        number of identical substrings has been captured.
1720    
1721    24. The way atomic groups are processed by pcre_exec() has been changed so that
1722        if they are repeated, backtracking one repetition now resets captured
1723        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1724        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1725        "aaa". Previously, it would have been "a". As part of this code
1726        refactoring, the way recursive calls are handled has also been changed.
1727    
1728    25. If an assertion condition captured any substrings, they were not passed
1729        back unless some other capturing happened later. For example, if
1730        (?(?=(a))a) was matched against "a", no capturing was returned.
1731    
1732    26. When studying a pattern that contained subroutine calls or assertions,
1733        the code for finding the minimum length of a possible match was handling
1734        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1735        group 1 called group 2 while simultaneously a separate group 2 called group
1736        1). A stack overflow occurred in this case. I have fixed this by limiting
1737        the recursion depth to 10.
1738    
1739    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1740        Fortmann. This supports explicit test numbers on the command line, and has
1741        argument validation and error reporting.
1742    
1743    28. An instance of \X with an unlimited repeat could fail if at any point the
1744        first character it looked at was a mark character.
1745    
1746    29. Some minor code refactoring concerning Unicode properties and scripts
1747        should reduce the stack requirement of match() slightly.
1748    
1749    30. Added the '=' option to pcretest to check the setting of unused capturing
1750        slots at the end of the pattern, which are documented as being -1, but are
1751        not included in the return count.
1752    
1753    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1754        compiled something random. Now it gives a compile-time error (as does
1755        Perl).
1756    
1757    32. A *MARK encountered during the processing of a positive assertion is now
1758        recorded and passed back (compatible with Perl).
1759    
1760    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1761        had alternative anchored branches, the search for a second match in a line
1762        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1763        matched the line "0102" twice. The same bug affected patterns that started
1764        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1765        twice.
1766    
1767    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1768        does, and because of capturing effects, quantifying parenthesized
1769        assertions may at times be useful. Quantifiers are now allowed for
1770        parenthesized assertions.
1771    
1772    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1773    
1774    36. \g was being checked for fancy things in a character class, when it should
1775        just be a literal "g".
1776    
1777    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1778        appearance of a nested POSIX class supersedes an apparent external class.
1779        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1780        unescaped square brackets may also appear as part of class names. For
1781        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1782        more like Perl. (But see 8.20/1 above.)
1783    
1784    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1785        was because it thought it was \N{name}, which is not supported).
1786    
1787    39. Add minix to OS list not supporting the -S option in pcretest.
1788    
1789    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1790        cannot analyze patterns in sufficient detail to catch mutual recursions
1791        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1792        subgroup is called recursively as a subpattern for a second time at the
1793        same position in the subject string. In previous releases this might have
1794        been caught by the recursion limit, or it might have run out of stack.
1795    
1796    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1797        happen only once. PCRE was, however incorrectly giving a compile time error
1798        "recursive call could loop indefinitely" because it cannot analyze the
1799        pattern in sufficient detail. The compile time test no longer happens when
1800        PCRE is compiling a conditional subpattern, but actual runaway loops are
1801        now caught at runtime (see 40 above).
1802    
1803    42. It seems that Perl allows any characters other than a closing parenthesis
1804        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1805        has been changed to be the same.
1806    
1807    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1808        as not to get warnings when autogen.sh is called. Also changed
1809        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1810    
1811    44. To help people who use pcregrep to scan files containing exceedingly long
1812        lines, the following changes have been made:
1813    
1814        (a) The default value of the buffer size parameter has been increased from
1815            8K to 20K. (The actual buffer used is three times this size.)
1816    
1817        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1818            PCRE is built.
1819    
1820        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1821            to be set at run time.
1822    
1823        (d) Numerical values in pcregrep options can be followed by K or M, for
1824            example --buffer-size=50K.
1825    
1826        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1827            given and the return code is set to 2.
1828    
1829    45. Add a pointer to the latest mark to the callout data block.
1830    
1831    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1832        partial match of an empty string instead of no match. This was specific to
1833        the use of ".".
1834    
1835    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1836        complete match instead of a partial match. This bug was dependent on both
1837        the PCRE_UTF8 and PCRE_DOTALL options being set.
1838    
1839    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1840        starting byte set, because \b was not being ignored.
1841    
1842    
1843    Version 8.12 15-Jan-2011
1844    ------------------------
1845    
1846    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1847        checks for such things as part of the documentation building process.
1848    
1849    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1850        --match-limit and --recursion-limit options (added for 8.11). In
1851        particular, this made one of the standard tests fail. (The integer value
1852        went into the wrong half of a long int.)
1853    
1854    3.  If the --colour option was given to pcregrep with -v (invert match), it
1855        did strange things, either producing crazy output, or crashing. It should,
1856        of course, ignore a request for colour when reporting lines that do not
1857        match.
1858    
1859    4.  Another pcregrep bug caused similar problems if --colour was specified with
1860        -M (multiline) and the pattern match finished with a line ending.
1861    
1862    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1863        matched in multiline mode, the following line was shown as part of the
1864        match. This seems wrong, so I have changed it.
1865    
1866    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1867        the check for further matches in the same line (so they could be coloured)
1868        to overrun the end of the current line. If another match was found, it was
1869        incorrectly shown (and then shown again when found in the next line).
1870    
1871    7.  If pcregrep was compiled under Windows, there was a reference to the
1872        function pcregrep_exit() before it was defined. I am assuming this was
1873        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1874        reported by a user. I've moved the definition above the reference.
1875    
1876    
1877    Version 8.11 10-Dec-2010
1878    ------------------------
1879    
1880    1.  (*THEN) was not working properly if there were untried alternatives prior
1881        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1882        backtracked to try for "b" instead of moving to the next alternative branch
1883        at the same level (in this case, to look for "c"). The Perl documentation
1884        is clear that when (*THEN) is backtracked onto, it goes to the "next
1885        alternative in the innermost enclosing group".
1886    
1887    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1888        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1889        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1890        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1891        (*THEN).
1892    
1893    3.  If \s appeared in a character class, it removed the VT character from
1894        the class, even if it had been included by some previous item, for example
1895        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1896        of \s, but is part of the POSIX "space" class.)
1897    
1898    4.  A partial match never returns an empty string (because you can always
1899        match an empty string at the end of the subject); however the checking for
1900        an empty string was starting at the "start of match" point. This has been
1901        changed to the "earliest inspected character" point, because the returned
1902        data for a partial match starts at this character. This means that, for
1903        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1904        (previously it gave "no match").
1905    
1906    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1907        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1908        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1909        has an implication that the given string is incomplete (because a partial
1910        match is preferred over a full match). For this reason, these items now
1911        give a partial match in this situation. [Aside: previously, the one case
1912        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1913        match rather than a full match, which was wrong by the old rules, but is
1914        now correct.]
1915    
1916    6.  There was a bug in the handling of #-introduced comments, recognized when
1917        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1918        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1919        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1920        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1921        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1922        places in pcre_compile().
1923    
1924    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1925        comments when looking ahead for named forward references to subpatterns,
1926        the only newline sequence it recognized was NL. It now handles newlines
1927        according to the set newline convention.
1928    
1929    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1930        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1931        cater for a lack of strerror(). These oversights have been fixed.
1932    
1933    9.  Added --match-limit and --recursion-limit to pcregrep.
1934    
1935    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1936    
1937    11. When the -o option was used, pcregrep was setting a return code of 1, even
1938        when matches were found, and --line-buffered was not being honoured.
1939    
1940    12. Added an optional parentheses number to the -o and --only-matching options
1941        of pcregrep.
1942    
1943    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1944        can match an empty string. The code to do it in pcretest and pcredemo
1945        needed fixing:
1946    
1947        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1948            only one byte after an empty string match just before CRLF (this case
1949            just got forgotten; "any" and "anycrlf" were OK).
1950    
1951        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1952            mode when an empty string match preceded an ASCII character followed by
1953            a non-ASCII character. (The code for advancing by one character rather
1954            than one byte was nonsense.)
1955    
1956        (c) The pcredemo.c sample program did not have any code at all to handle
1957            the cases when CRLF is a valid newline sequence.
1958    
1959    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1960        as a starting offset was within the subject string. There is now a new
1961        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1962        negative or greater than the length of the string. In order to test this,
1963        pcretest is extended to allow the setting of negative starting offsets.
1964    
1965    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1966        starting offset points to the beginning of a UTF-8 character was
1967        unnecessarily clumsy. I tidied it up.
1968    
1969    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1970        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1971    
1972    17. Nobody had reported that the --include_dir option, which was added in
1973        release 7.7 should have been called --include-dir (hyphen, not underscore)
1974        for compatibility with GNU grep. I have changed it to --include-dir, but
1975        left --include_dir as an undocumented synonym, and the same for
1976        --exclude-dir, though that is not available in GNU grep, at least as of
1977        release 2.5.4.
1978    
1979    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1980        characters from a string of bytes) have been redefined so as not to use
1981        loops, in order to improve performance in some environments. At the same
1982        time, I abstracted some of the common code into auxiliary macros to save
1983        repetition (this should not affect the compiled code).
1984    
1985    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1986        compile-time error is now given if \c is not followed by an ASCII
1987        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1988        different, and any byte value is allowed.)
1989    
1990    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1991        START_OPTIMIZE option, which is now allowed at compile time - but just
1992        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1993        to pcregrep and other applications that have no direct access to PCRE
1994        options. The new /Y option in pcretest sets this option when calling
1995        pcre_compile().
1996    
1997    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1998        back references. Groups containing recursive back references were forced to
1999        be atomic by that change, but in the case of named groups, the amount of
2000        memory required was incorrectly computed, leading to "Failed: internal
2001        error: code overflow". This has been fixed.
2002    
2003    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2004        pcretest.c, to avoid build problems in some Borland environments.
2005    
2006    
2007    Version 8.10 25-Jun-2010
2008    ------------------------
2009    
2010    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2011        THEN.
2012    
2013    2.  (*ACCEPT) was not working when inside an atomic group.
2014    
2015    3.  Inside a character class, \B is treated as a literal by default, but
2016        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2017        causes the error). The code is unchanged, but I tidied the documentation.
2018    
2019    4.  Inside a character class, PCRE always treated \R and \X as literals,
2020        whereas Perl faults them if its -w option is set. I have changed PCRE so
2021        that it faults them when PCRE_EXTRA is set.
2022    
2023    5.  Added support for \N, which always matches any character other than
2024        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2025    
2026    6.  When compiling pcregrep with newer versions of gcc which may have
2027        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2028        declared with attribute warn_unused_result" were given. Just casting the
2029        result to (void) does not stop the warnings; a more elaborate fudge is
2030        needed. I've used a macro to implement this.
2031    
2032    7.  Minor change to pcretest.c to avoid a compiler warning.
2033    
2034    8.  Added four artifical Unicode properties to help with an option to make
2035        \s etc use properties (see next item). The new properties are: Xan
2036        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2037    
2038    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2039        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2040        this option. Modified pcretest to add /W to test this facility. Added
2041        REG_UCP to make it available via the POSIX interface.
2042    
2043    10. Added --line-buffered to pcregrep.
2044    
2045    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2046        studied, and the match started with a letter with a code point greater than
2047        127 whose first byte was different to the first byte of the other case of
2048        the letter, the other case of this starting letter was not recognized
2049        (#976).
2050    
2051    12. If a pattern that was studied started with a repeated Unicode property
2052        test, for example, \p{Nd}+, there was the theoretical possibility of
2053        setting up an incorrect bitmap of starting bytes, but fortunately it could
2054        not have actually happened in practice until change 8 above was made (it
2055        added property types that matched character-matching opcodes).
2056    
2057    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2058        possible starting bytes for non-anchored patterns.
2059    
2060    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2061        \R, and also a number of cases that involve Unicode properties, both
2062        explicit and implicit when PCRE_UCP is set.
2063    
2064    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2065        input, it could crash or give wrong results if characters with values
2066        greater than 0xc0 were present in the subject string. (Detail: it assumed
2067        UTF-8 input when processing these items.)
2068    
2069    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2070        size_t is 64-bit (#991).
2071    
2072    17. Added a check for running out of memory when PCRE is compiled with
2073        --disable-stack-for-recursion (#990).
2074    
2075    18. If the last data line in a file for pcretest does not have a newline on
2076        the end, a newline was missing in the output.
2077    
2078    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2079        less than 128) in its various bitmaps. However, there is a facility for
2080        generating tables according to the current locale when PCRE is compiled. It
2081        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2082        characters, are recognized by isspace() and therefore were getting set in
2083        these tables, and indeed these tables seem to approximate to ISO 8859. This
2084        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2085        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2086        which of course cannot start UTF-8 characters. I have changed the code so
2087        that only real ASCII characters (less than 128) and the correct starting
2088        bytes for UTF-8 encodings are set for characters greater than 127 when in
2089        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2090        altogether.)
2091    
2092    20. Added the /T option to pcretest so as to be able to run tests with non-
2093        standard character tables, thus making it possible to include the tests
2094        used for 19 above in the standard set of tests.
2095    
2096    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2097        reference to a subpattern the other side of a comment that contains an
2098        opening parenthesis caused either an internal compiling error, or a
2099        reference to the wrong subpattern.
2100    
2101    
2102    Version 8.02 19-Mar-2010
2103    ------------------------
2104    
2105    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2106    
2107    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2108        configured.
2109    
2110    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2111        original author of that file, following a query about its status.
2112    
2113    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2114        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2115    
2116    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2117        quantifier applied to a forward-referencing subroutine call, could compile
2118        incorrect code or give the error "internal error: previously-checked
2119        referenced subpattern not found".
2120    
2121    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2122        variables to point to external functions. For these systems, therefore,
2123        pcre_malloc etc. are now initialized to local functions that call the
2124        relevant global functions.
2125    
2126    7.  There were two entries missing in the vectors called coptable and poptable
2127        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2128        I've fixed the data, and added a kludgy way of testing at compile time that
2129        the lengths are correct (equal to the number of opcodes).
2130    
2131    8.  Following on from 7, I added a similar kludge to check the length of the
2132        eint vector in pcreposix.c.
2133    
2134    9.  Error texts for pcre_compile() are held as one long string to avoid too
2135        much relocation at load time. To find a text, the string is searched,
2136        counting zeros. There was no check for running off the end of the string,
2137        which could happen if a new error number was added without updating the
2138        string.
2139    
2140    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2141    
2142    11. \K was not working if it appeared in an atomic group or in a group that
2143        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2144        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2145        the assertion is positive, but not if it is negative.
2146    
2147    12. Change 11 fortuitously reduced the size of the stack frame used in the
2148        "match()" function of pcre_exec.c by one pointer. Forthcoming
2149        implementation of support for (*MARK) will need an extra pointer on the
2150        stack; I have reserved it now, so that the stack frame size does not
2151        decrease.
2152    
2153    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2154        item in branch that calls a recursion is a subroutine call - as in the
2155        second branch in the above example - was incorrectly given the compile-
2156        time error "recursive call could loop indefinitely" because pcre_compile()
2157        was not correctly checking the subroutine for matching a non-empty string.
2158    
2159    14. The checks for overrunning compiling workspace could trigger after an
2160        overrun had occurred. This is a "should never occur" error, but it can be
2161        triggered by pathological patterns such as hundreds of nested parentheses.
2162        The checks now trigger 100 bytes before the end of the workspace.
2163    
2164    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2165    
2166    
2167    Version 8.01 19-Jan-2010
2168    ------------------------
2169    
2170    1.  If a pattern contained a conditional subpattern with only one branch (in
2171        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2172        computed the wrong minimum data length (which is of course zero for such
2173        subpatterns). This could cause incorrect "no match" results.
2174    
2175    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2176        the pattern is reset in the first branch, pcre_compile() failed with
2177        "internal error: code overflow at offset...". This happened only when
2178        the reset was to the original external option setting. (An optimization
2179        abstracts leading options settings into an external setting, which was the
2180        cause of this.)
2181    
2182    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2183        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2184        assertion pattern did not match (meaning that the assertion was true), it
2185        was incorrectly treated as false if the SKIP had been reached during the
2186        matching. This also applied to assertions used as conditions.
2187    
2188    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2189        assertion subpattern, including such a pattern used as a condition,
2190        unpredictable results occurred, instead of the error return
2191        PCRE_ERROR_DFA_UITEM.
2192    
2193    5.  The C++ GlobalReplace function was not working like Perl for the special
2194        situation when an empty string is matched. It now does the fancy magic
2195        stuff that is necessary.
2196    
2197    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2198        removed. (These were left over from very, very early versions of PCRE.)
2199    
2200    7.  Some cosmetic changes to the code to make life easier when compiling it
2201        as part of something else:
2202    
2203        (a) Change DEBUG to PCRE_DEBUG.
2204    
2205        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2206            called "current" as "current_branch", to prevent a collision with the
2207            Linux macro when compiled as a kernel module.
2208    
2209        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2210            prevent a collision with the Linux macro when compiled as a kernel
2211            module.
2212    
2213    8.  In pcre_compile() there are some checks for integer overflows that used to
2214        cast potentially large values to (double). This has been changed to that
2215        when building, a check for int64_t is made, and if it is found, it is used
2216        instead, thus avoiding the use of floating point arithmetic. (There is no
2217        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2218        double.
2219    
2220    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2221        2005 (difference between two addresses compared to an unsigned value).
2222    
2223    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2224        custom one, because of the following reported problem in Windows:
2225    
2226          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2227              under Win32.
2228          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2229              therefore missing the function definition.
2230          - The compiler thus generates a "C" signature for the test function.
2231          - The linker fails to find the "C" function.
2232          - PCRE fails to configure if asked to do so against libbz2.
2233    
2234    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2235        messages were output:
2236    
2237          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2238          rerunning libtoolize, to keep the correct libtool macros in-tree.
2239          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2240    
2241        I have done both of these things.
2242    
2243    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2244        most of the time, it *can* run out if it is given a pattern that contains a
2245        runaway infinite recursion. I updated the discussion in the pcrestack man
2246        page.
2247    
2248    13. Now that we have gone to the x.xx style of version numbers, the minor
2249        version may start with zero. Using 08 or 09 is a bad idea because users
2250        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2251        interpreted as invalid octal numbers. I've updated the previous comment in
2252        configure.ac, and also added a check that gives an error if 08 or 09 are
2253        used.
2254    
2255    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2256        causing partial matching to fail when the end of the subject matched \W
2257        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2258    
2259    15. There were some discrepancies between the declarations in pcre_internal.h
2260        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2261        their definitions. The declarations used "const uschar *" and the
2262        definitions used USPTR. Even though USPTR is normally defined as "const
2263        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2264        reported that: "This difference in casting confuses some C++ compilers, for
2265        example, SunCC recognizes above declarations as different functions and
2266        generates broken code for hbpcre." I have changed the declarations to use
2267        USPTR.
2268    
2269    16. GNU libtool is named differently on some systems. The autogen.sh script now
2270        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2271        (FreeBSD).
2272    
2273    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2274        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2275        comment: "Figure out how to create a longlong from a string: strtoll and
2276        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2277        instance, but it only takes 2 args instead of 3!"
2278    
2279    18. A subtle bug concerned with back references has been fixed by a change of
2280        specification, with a corresponding code fix. A pattern such as
2281        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2282        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2283        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2284        same bug. Such groups have to be quantified to be useful, or contained
2285        inside another quantified group. (If there's no repetition, the reference
2286        can never match.) The problem arises because, having left the group and
2287        moved on to the rest of the pattern, a later failure that backtracks into
2288        the group uses the captured value from the final iteration of the group
2289        rather than the correct earlier one. I have fixed this in PCRE by forcing
2290        any group that contains a reference to itself to be an atomic group; that
2291        is, there cannot be any backtracking into it once it has completed. This is
2292        similar to recursive and subroutine calls.
2293    
2294    
2295    Version 8.00 19-Oct-09
2296    ----------------------
2297    
2298    1.  The table for translating pcre_compile() error codes into POSIX error codes
2299        was out-of-date, and there was no check on the pcre_compile() error code
2300        being within the table. This could lead to an OK return being given in
2301        error.
2302    
2303    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2304        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2305        in a Windows environment.
2306    
2307    3.  The pcregrep --count option prints the count for each file even when it is
2308        zero, as does GNU grep. However, pcregrep was also printing all files when
2309        --files-with-matches was added. Now, when both options are given, it prints
2310        counts only for those files that have at least one match. (GNU grep just
2311        prints the file name in this circumstance, but including the count seems
2312        more useful - otherwise, why use --count?) Also ensured that the
2313        combination -clh just lists non-zero counts, with no names.
2314    
2315    4.  The long form of the pcregrep -F option was incorrectly implemented as
2316        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2317        but it seems right to fix it, and I didn't think it was worth preserving
2318        the old behaviour.
2319    
2320    5.  The command line items --regex=pattern and --regexp=pattern were not
2321        recognized by pcregrep, which required --regex pattern or --regexp pattern
2322        (with a space rather than an '='). The man page documented the '=' forms,
2323        which are compatible with GNU grep; these now work.
2324    
2325    6.  No libpcreposix.pc file was created for pkg-config; there was just
2326        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2327    
2328    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2329        when UCP support is not needed, by modifying the Python script that
2330        generates it from Unicode data files. This should not matter if the module
2331        is correctly used as a library, but I received one complaint about 50K of
2332        unwanted data. My guess is that the person linked everything into his
2333        program rather than using a library. Anyway, it does no harm.
2334    
2335    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2336        was a minimum greater than 1 for a wide character in a possessive
2337        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2338        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2339        character. Chaos in the form of incorrect output or a compiling loop could
2340        result.
2341    
2342    9.  The restrictions on what a pattern can contain when partial matching is
2343        requested for pcre_exec() have been removed. All patterns can now be
2344        partially matched by this function. In addition, if there are at least two
2345        slots in the offset vector, the offset of the earliest inspected character
2346        for the match and the offset of the end of the subject are set in them when
2347        PCRE_ERROR_PARTIAL is returned.
2348    
2349    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2350        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2351        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2352        and may be more useful for multi-segment matching.
2353    
2354    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2355        used to be given if ever the end of the subject was reached; now it is
2356        given only if matching could not proceed because another character was
2357        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2358        string "Z", which now yields "no match" instead of "partial match". In the
2359        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2360        final character ended with (*FAIL).
2361    
2362    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2363        if the pattern had a "must contain" character that was already found in the
2364        earlier partial match, unless partial matching was again requested. For
2365        example, with the pattern /dog.(body)?/, the "must contain" character is
2366        "g". If the first part-match was for the string "dog", restarting with
2367        "sbody" failed. This bug has been fixed.
2368    
2369    13. The string returned by pcre_dfa_exec() after a partial match has been
2370        changed so that it starts at the first inspected character rather than the
2371        first character of the match. This makes a difference only if the pattern
2372        starts with a lookbehind assertion or \b or \B (\K is not supported by
2373        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2374        matching functions compatible, and I think it's the right thing to do.
2375    
2376    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2377        so that the demonstration program is easily available in environments where
2378        PCRE has not been installed from source.
2379    
2380    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2381        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2382        library.
2383    
2384    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2385        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2386        is not the first non-POSIX option to be added. Clearly some people find
2387        these options useful.
2388    
2389    17. If a caller to the POSIX matching function regexec() passes a non-zero
2390        value for nmatch with a NULL value for pmatch, the value of
2391        nmatch is forced to zero.
2392    
2393    18. RunGrepTest did not have a test for the availability of the -u option of
2394        the diff command, as RunTest does. It now checks in the same way as
2395        RunTest, and also checks for the -b option.
2396    
2397    19. If an odd number of negated classes containing just a single character
2398        interposed, within parentheses, between a forward reference to a named
2399        subpattern and the definition of the subpattern, compilation crashed with
2400        an internal error, complaining that it could not find the referenced
2401        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2402        [The bug was that it was starting one character too far in when skipping
2403        over the character class, thus treating the ] as data rather than
2404        terminating the class. This meant it could skip too much.]
2405    
2406    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2407        /g option in pcretest when the pattern contains \K, which makes it possible
2408        to have an empty string match not at the start, even when the pattern is
2409        anchored. Updated pcretest and pcredemo to use this option.
2410    
2411    21. If the maximum number of capturing subpatterns in a recursion was greater
2412        than the maximum at the outer level, the higher number was returned, but
2413        with unset values at the outer level. The correct (outer level) value is
2414        now given.
2415    
2416    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2417        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2418        make it do so. The string so far is captured, making this feature
2419        compatible with Perl.
2420    
2421    23. The tests have been re-organized, adding tests 11 and 12, to make it
2422        possible to check the Perl 5.10 features against Perl 5.10.
2423    
2424    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2425        pattern matches a fixed length string. PCRE did not allow this; now it
2426        does. Neither allows recursion.
2427    
2428    25. I finally figured out how to implement a request to provide the minimum
2429        length of subject string that was needed in order to match a given pattern.
2430        (It was back references and recursion that I had previously got hung up
2431        on.) This code has now been added to pcre_study(); it finds a lower bound
2432        to the length of subject needed. It is not necessarily the greatest lower
2433        bound, but using it to avoid searching strings that are too short does give
2434        some useful speed-ups. The value is available to calling programs via
2435        pcre_fullinfo().
2436    
2437    26. While implementing 25, I discovered to my embarrassment that pcretest had
2438        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2439        study optimizations had never been tested with that matching function.
2440        Oops. What is worse, even when it was passed study data, there was a bug in
2441        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2442        were also very few tests of studied patterns with pcre_dfa_exec().
2443    
2444    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2445        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2446        on the other side of the coin, they are no longer allowed to have different
2447        names, because these cannot be distinguished in PCRE, and this has caused
2448        confusion. (This is a difference from Perl.)
2449    
2450    28. When duplicate subpattern names are present (necessarily with different
2451        numbers, as required by 27 above), and a test is made by name in a
2452        conditional pattern, either for a subpattern having been matched, or for
2453        recursion in such a pattern, all the associated numbered subpatterns are
2454        tested, and the overall condition is true if the condition is true for any
2455        one of them. This is the way Perl works, and is also more like the way
2456        testing by number works.
2457    
2458    
2459    Version 7.9 11-Apr-09
2460    ---------------------
2461    
2462    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2463        (pcretest), all targets were linked against these libraries. This included
2464        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2465        libraries. This caused unwanted dependencies to be created. This problem
2466        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2467        pcretest is linked with readline.
2468    
2469    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2470        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2471        moved outside it again, because FALSE and TRUE are already defined in AIX,
2472        but BOOL is not.
2473    
2474    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2475        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2476    
2477    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2478        hyphens) following file names and line numbers when outputting matching
2479        lines. This is not true; no spaces are inserted. I have also clarified the
2480        wording for the --colour (or --color) option.
2481    
2482    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2483        was not coloured; this is different to GNU grep, so I have changed it to be
2484        the same.
2485    
2486    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2487        each matching line was coloured. Now it goes on to look for further matches
2488        of any of the test patterns, which is the same behaviour as GNU grep.
2489    
2490    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2491        doesn't make sense to accept an empty string match in pcregrep, so I have
2492        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2493        seems to be how GNU grep behaves. [But see later change 40 for release
2494        8.33.]
2495    
2496    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2497        start or after a newline", because the conditional assertion was not being
2498        correctly handled. The rule now is that both the assertion and what follows
2499        in the first alternative must satisfy the test.
2500    
2501    9.  If auto-callout was enabled in a pattern with a conditional group whose
2502        condition was an assertion, PCRE could crash during matching, both with
2503        pcre_exec() and pcre_dfa_exec().
2504    
2505    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2506        used for matching.
2507    
2508    11. Unicode property support in character classes was not working for
2509        characters (bytes) greater than 127 when not in UTF-8 mode.
2510    
2511    12. Added the -M command line option to pcretest.
2512    
2513    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2514    
2515    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2516    
2517    16. Added comments and documentation about mis-use of no_arg in the C++
2518        wrapper.
2519    
2520    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2521        from Martin Jerabek that uses macro names for all relevant character and
2522        string constants.
2523    
2524    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2525        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2526        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2527        these, but not everybody uses configure.
2528    
2529    19. A conditional group that had only one branch was not being correctly
2530        recognized as an item that could match an empty string. This meant that an
2531        enclosing group might also not be so recognized, causing infinite looping
2532        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2533        with the subject "ab", where knowledge that the repeated group can match
2534        nothing is needed in order to break the loop.
2535    
2536    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2537        exec(), but without supplying a callout function, matching went wrong.
2538    
2539    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2540        leak if the size of the offset vector was greater than 30. When the vector
2541        is smaller, the saved offsets during recursion go onto a local stack
2542        vector, but for larger vectors malloc() is used. It was failing to free
2543        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2544        error, in fact).
2545    
2546    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2547        heapframe that is used only when UTF-8 support is enabled. This caused no
2548        problem, but was untidy.
2549    
2550    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2551        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2552        included within another project.
2553    
2554    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2555        slightly modified by me:
2556    
2557          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2558              not building pcregrep.
2559    
2560          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2561              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2562    
2563    25. Forward references, both numeric and by name, in patterns that made use of
2564        duplicate group numbers, could behave incorrectly or give incorrect errors,
2565        because when scanning forward to find the reference group, PCRE was not
2566        taking into account the duplicate group numbers. A pattern such as
2567        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2568    
2569    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2570        the feature of a custom pointer more persuasive (as requested by a user).
2571    
2572    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2573        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2574        pre-defined.
2575    
2576    28. Added support for (*UTF8) at the start of a pattern.
2577    
2578    29. Arrange for flags added by the "release type" setting in CMake to be shown
2579        in the configuration summary.
2580    
2581    
2582    Version 7.8 05-Sep-08
2583    ---------------------
2584    
2585    1.  Replaced UCP searching code with optimized version as implemented for Ad
2586        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2587        stage table and inline lookup instead of a function, giving speed ups of 2
2588        to 5 times on some simple patterns that I tested. Permission was given to
2589        distribute the MultiStage2.py script that generates the tables (it's not in
2590        the tarball, but is in the Subversion repository).
2591    
2592    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2593        scripts.
2594    
2595    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2596        a group with a zero qualifier. The result of the study could be incorrect,
2597        or the function might crash, depending on the pattern.
2598    
2599    4.  Caseless matching was not working for non-ASCII characters in back
2600        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2601        It now works when Unicode Property Support is available.
2602    
2603    5.  In pcretest, an escape such as \x{de} in the data was always generating
2604        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2605        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2606        truncation.
2607    
2608    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2609    
2610    7.  Added two (int) casts to pcregrep when printing the difference of two
2611        pointers, in case they are 64-bit values.
2612    
2613    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2614        test 2 if it fails.
2615    
2616    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2617        and a #define of that name to empty if it is not externally set. This is to
2618        allow users of MSVC to set it if necessary.
2619    
2620    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2621        the convenience functions in the pcre_get.c source file.
2622    
2623    11. An option change at the start of a pattern that had top-level alternatives
2624        could cause overwriting and/or a crash. This command provoked a crash in
2625        some environments:
2626    
2627          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2628    
2629        This potential security problem was recorded as CVE-2008-2371.
2630    
2631    12. For a pattern where the match had to start at the beginning or immediately
2632        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2633        pcre_dfa_exec() could read past the end of the passed subject if there was
2634        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2635        pcretest so that it places the subject at the end of its malloc-ed buffer.
2636    
2637    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2638        exec() might read past the end of the data buffer in UTF-8 mode.
2639    
2640    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2641        the data contained the byte 0x85 as part of a UTF-8 character within its
2642        first line. This applied both to normal and DFA matching.
2643    
2644    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2645        /^[^d]*?$/8 failed to match "abc".
2646    
2647    16. Added a missing copyright notice to pcrecpp_internal.h.
2648    
2649    17. Make it more clear in the documentation that values returned from
2650        pcre_exec() in ovector are byte offsets, not character counts.
2651    
2652    18. Tidied a few places to stop certain compilers from issuing warnings.
2653    
2654    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2655        supplied by Stefan Weber. I made a further small update for 7.8 because
2656        there is a change of source arrangements: the pcre_searchfuncs.c module is
2657        replaced by pcre_ucd.c.
2658    
2659    
2660    Version 7.7 07-May-08
2661    ---------------------
2662    
2663    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2664        a string to a long long, pretend we don't even have a long long." This is
2665        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2666    
2667    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2668        pre-7.6 versions, which defined a global no_arg variable instead of putting
2669        it in the RE class. (See also #8 below.)
2670    
2671    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2672        Lopes.
2673    
2674    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2675    
2676        (1) The include/exclude patterns were being applied to the whole pathnames
2677            of files, instead of just to the final components.
2678    
2679        (2) If there was more than one level of directory, the subdirectories were
2680            skipped unless they satisfied the include/exclude conditions. This is
2681            inconsistent with GNU grep (and could even be seen as contrary to the
2682            pcregrep specification - which I improved to make it absolutely clear).
2683            The action now is always to scan all levels of directory, and just
2684            apply the include/exclude patterns to regular files.
2685    
2686    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2687        --exclude_dir in the tests to avoid scanning .svn directories.
2688    
2689    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2690        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2691        doesn't support NULs in patterns.
2692    
2693    7.  Added some missing "const"s to declarations of static tables in
2694        pcre_compile.c and pcre_dfa_exec.c.
2695    
2696    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2697        caused by fix #2  above. (Subsequently also a second patch to fix the
2698        first patch. And a third patch - this was a messy problem.)
2699    
2700    9.  Applied Craig's patch to remove the use of push_back().
2701    
2702    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2703        matching function regexec().
2704    
2705    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2706        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2707        references. PCRE supports relative numbers with this syntax (I don't think
2708        Oniguruma does).
2709    
2710    12. Previously, a group with a zero repeat such as (...){0} was completely
2711        omitted from the compiled regex. However, this means that if the group
2712        was called as a subroutine from elsewhere in the pattern, things went wrong
2713        (an internal error was given). Such groups are now left in the compiled
2714        pattern, with a new opcode that causes them to be skipped at execution
2715        time.
2716    
2717    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2718        to the way PCRE behaves:
2719    
2720        (a) A lone ] character is dis-allowed (Perl treats it as data).
2721    
2722        (b) A back reference to an unmatched subpattern matches an empty string
2723            (Perl fails the current match path).
2724    
2725        (c) A data ] in a character class must be notated as \] because if the
2726            first data character in a class is ], it defines an empty class. (In
2727            Perl it is not possible to have an empty class.) The empty class []
2728            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2729            The negative empty class [^] matches any one character, independently
2730            of the DOTALL setting.
2731    
2732    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2733        non-existent subpattern following a character class starting with ']' and
2734        containing () gave an internal compiling error instead of "reference to
2735        non-existent subpattern". Fortunately, when the pattern did exist, the
2736        compiled code was correct. (When scanning forwards to check for the
2737        existence of the subpattern, it was treating the data ']' as terminating
2738        the class, so got the count wrong. When actually compiling, the reference
2739        was subsequently set up correctly.)
2740    
2741    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2742        it was being rejected as not supported by pcre_dfa_exec(), even though
2743        other assertions are supported. I have made pcre_dfa_exec() support
2744        (*FAIL).
2745    
2746    16. The implementation of 13c above involved the invention of a new opcode,
2747        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2748        cannot be changed at match time, I realized I could make a small
2749        improvement to matching performance by compiling OP_ALLANY instead of
2750        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2751        on the OP_ANY path.
2752    
2753    17. Compiling pcretest on Windows with readline support failed without the
2754        following two fixes: (1) Make the unistd.h include conditional on
2755        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2756    
2757    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2758        ncurses library to be included for pcretest when ReadLine support is
2759        requested, but also to allow for it to be overridden. This patch came from
2760        Daniel Bergström.
2761    
2762    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2763        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2764        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2765        spotting this.
2766    
2767    
2768    Version 7.6 28-Jan-08
2769    ---------------------
2770    
2771    1.  A character class containing a very large number of characters with
2772        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2773        overflow.
2774    
2775    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2776        HAVE_LONG_LONG is not defined.
2777    
2778    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2779        bring them up to date and include new features. This patch includes:
2780    
2781        - Fixed PH's badly added libz and libbz2 support.
2782        - Fixed a problem with static linking.
2783        - Added pcredemo. [But later removed - see 7 below.]
2784        - Fixed dftables problem and added an option.
2785        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2786            HAVE_LONG_LONG.
2787        - Added readline support for pcretest.
2788        - Added an listing of the option settings after cmake has run.
2789    
2790    4.  A user submitted a patch to Makefile that makes it easy to create
2791        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2792        Makefile.am that cause it to include this special target, without
2793        affecting anything else. Note that the same mingw target plus all
2794        the other distribution libraries and programs are now supported
2795        when configuring with CMake (see 6 below) instead of with
2796        Configure/Make.
2797    
2798    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2799        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2800        exported in the Windows port". It has not yet been confirmed that the patch
2801        solves the problem, but it does no harm.
2802    
2803    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2804        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2805        with CMake, and also correct the comment about stack recursion.
2806    
2807    7.  Remove the automatic building of pcredemo from the ./configure system and
2808        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2809        of a program that users should build themselves after PCRE is installed, so
2810        building it automatically is not really right. What is more, it gave
2811        trouble in some build environments.
2812    
2813    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2814    
2815    
2816    Version 7.5 10-Jan-08
2817    ---------------------
2818    
2819    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2820        values in parens when parsing an RE using the C++ wrapper."
2821    
2822    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2823        Characters greater than 255 were excluded from the class instead of being
2824        included.
2825    
2826    3.  The same bug as (2) above applied to negated POSIX classes such as
2827        [:^space:].
2828    
2829    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2830        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2831        I have changed it.
2832    
2833    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2834        first named subpattern) and a construct such as (?&a) would reference the
2835        first named subpattern whose name started with "a" (in other words, the
2836        length check was missing). Both these problems are fixed. "Subpattern name
2837        expected" is now given for (?&) (a zero-length name), and this patch also
2838        makes it give the same error for \k'' (previously it complained that that
2839        was a reference to a non-existent subpattern).
2840    
2841    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2842        this is right because (?- can be followed by option settings as well as by
2843        digits. I have, however, made the messages clearer.
2844    
2845    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2846        than the number used in the conditional) now cause a compile-time error.
2847        This is actually not compatible with Perl, which accepts such patterns, but
2848        treats the conditional as always being FALSE (as PCRE used to), but it
2849        seems to me that giving a diagnostic is better.
2850    
2851    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2852        and messages.
2853    
2854    9.  Fix two occurrences of "backslash" in comments that should have been
2855        "backspace".
2856    
2857    10. Remove two redundant lines of code that can never be obeyed (their function
2858        was moved elsewhere).
2859    
2860    11. The program that makes PCRE's Unicode character property table had a bug
2861        which caused it to generate incorrect table entries for sequences of
2862        characters that have the same character type, but are in different scripts.
2863        It amalgamated them into a single range, with the script of the first of
2864        them. In other words, some characters were in the wrong script. There were
2865        thirteen such cases, affecting characters in the following ranges:
2866    
2867          U+002b0 - U+002c1
2868          U+0060c - U+0060d
2869          U+0061e - U+00612
2870          U+0064b - U+0065e
2871          U+0074d - U+0076d
2872          U+01800 - U+01805
2873          U+01d00 - U+01d77
2874          U+01d9b - U+01dbf
2875          U+0200b - U+0200f
2876          U+030fc - U+030fe
2877          U+03260 - U+0327f
2878          U+0fb46 - U+0fbb1
2879          U+10450 - U+1049d
2880    
2881    12. The -o option (show only the matching part of a line) for pcregrep was not
2882        compatible with GNU grep in that, if there was more than one match in a
2883        line, it showed only the first of them. It now behaves in the same way as
2884        GNU grep.
2885    
2886    13. If the -o and -v options were combined for pcregrep, it printed a blank
2887        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2888        does the same. The return code can be used to tell if there were any
2889        non-matching lines.
2890    
2891    14. Added --file-offsets and --line-offsets to pcregrep.
2892    
2893    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2894        infinitely looping recursion. The bug was that positive lookaheads were not
2895        being skipped when checking for a possible empty match (negative lookaheads
2896        and both kinds of lookbehind were skipped).
2897    
2898    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2899        inclusion of <windows.h> to before rather than after the definition of
2900        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2901    
2902    17. Specifying a possessive quantifier with a specific limit for a Unicode
2903        character property caused pcre_compile() to compile bad code, which led at
2904        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2905        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2906        caused the error; without that there was no problem.
2907    
2908    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2909    
2910    19. Added --enable-pcretest-libreadline.
2911    
2912    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2913        RE::GlobalReplace(). As a result, the number of replacements returned was
2914        double what it should be. I removed one of the increments, but Craig sent a
2915        later patch that removed the other one (the right fix) and added unit tests
2916        that check the return values (which was not done before).
2917    
2918    21. Several CMake things:
2919    
2920        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2921            the names libpcre and libpcreposix, not just pcre and pcreposix.
2922    
2923        (2) The above change means that pcretest and pcregrep are now correctly
2924            linked with the newly-built libraries, not previously installed ones.
2925    
2926        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2927    
2928    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2929        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2930        UTF-8 newline character). The key issue is that the pattern starts .*;
2931        this means that the match must be either at the beginning, or after a
2932        newline. The bug was in the code for advancing after a failed match and
2933        checking that the new position followed a newline. It was not taking
2934        account of UTF-8 characters correctly.
2935    
2936    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2937        character classes. PCRE was not treating the sequence [:...:] as a
2938        character class unless the ... were all letters. Perl, however, seems to
2939        allow any characters between [: and :], though of course it rejects as
2940        unknown any "names" that contain non-letters, because all the known class
2941        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2942        for example, whereas PCRE did not - it did not recognize a POSIX character
2943        class. This seemed a bit dangerous, so the code has been changed to be
2944        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2945        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2946        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2947        Perl does, and where it didn't before.
2948    
2949    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2950        Windows environments %n is disabled by default.
2951    
2952    
2953    Version 7.4 21-Sep-07
2954    ---------------------
2955    
2956    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2957        means that a class such as [\s] counted as "explicit reference to CR or
2958        LF". That isn't really right - the whole point of the change was to try to
2959        help when there was an actual mention of one of the two characters. So now
2960        the change happens only if \r or \n (or a literal CR or LF) character is
2961        encountered.
2962    
2963    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2964        of both had grown to the point where there were only 3 bits left.
2965        Fortunately, there was spare space in the data structure, and so I have
2966        moved the internal flags into a new 16-bit field to free up more option
2967        bits.
2968    
2969    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2970        but did not set the internal JCHANGED flag - either of these is enough to
2971        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2972        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2973        start sets both bits.
2974    
2975    4.  Added options (at build time, compile time, exec time) to change \R from
2976        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2977    
2978    5.  doc/pcresyntax.html was missing from the distribution.
2979    
2980    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2981        compatibility, even though it is no longer used.
2982    
2983    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2984        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2985        windows.h file is present (where different names are used). [This was
2986        reversed later after testing - see 16 below.]
2987    
2988    8.  Changed all #include <config.h> to #include "config.h". There were also
2989        some further <pcre.h> cases that I changed to "pcre.h".
2990    
2991    9.  When pcregrep was used with the --colour option, it missed the line ending
2992        sequence off the lines that it output.
2993    
2994    10. It was pointed out to me that arrays of string pointers cause lots of
2995        relocations when a shared library is dynamically loaded. A technique of
2996        using a single long string with a table of offsets can drastically reduce
2997        these. I have refactored PCRE in four places to do this. The result is
2998        dramatic:
2999    
3000          Originally:                          290
3001          After changing UCP table:            187
3002          After changing error message table:   43
3003          After changing table of "verbs"       36
3004          After changing table of Posix names   22
3005    
3006        Thanks to the folks working on Gregex for glib for this insight.
3007    
3008    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3009        unicode-properties was also set.
3010    
3011    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3012    
3013    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3014        checked only for CRLF.
3015    
3016    14. Added casts to pcretest.c to avoid compiler warnings.
3017    
3018    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3019    
3020    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3021        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3022        entirely. This removes changes made in 7 above.
3023    
3024    17. The CMake files have been updated, and there is now more information about
3025        building with CMake in the NON-UNIX-USE document.
3026    
3027    
3028    Version 7.3 28-Aug-07
3029    ---------------------
3030    
3031     1. In the rejigging of the build system that eventually resulted in 7.1, the
3032        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3033        brackets there is not right, since it causes compilers to look for an
3034        installed pcre.h, not the version that is in the source that is being
3035        compiled (which of course may be different). I have changed it back to:
3036    
3037          #include "pcre.h"
3038    
3039        I have a vague recollection that the change was concerned with compiling in
3040        different directories, but in the new build system, that is taken care of
3041        by the VPATH setting the Makefile.
3042    
3043     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3044        when the subject happened to end in the byte 0x85 (e.g. if the last
3045        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3046        characters but of course it shouldn't be taken as a newline when it is part
3047        of another character. The bug was that, for an unlimited repeat of . in
3048        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3049        characters when looking for a newline.
3050    
3051     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3052    
3053     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3054        in debug output.
3055    
3056     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3057        long printing in the pcrecpp unittest when running under MinGW.
3058    
3059     6. ESC_K was left out of the EBCDIC table.
3060    
3061     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3062        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3063        limit also applies to "virtual nesting" when a pattern is recursive, and in
3064        this case 1000 isn't so big. I have been able to remove this limit at the
3065        expense of backing off one optimization in certain circumstances. Normally,
3066        when pcre_exec() would call its internal match() function recursively and
3067        immediately return the result unconditionally, it uses a "tail recursion"
3068        feature to save stack. However, when a subpattern that can match an empty
3069        string has an unlimited repetition quantifier, it no longer makes this
3070        optimization. That gives it a stack frame in which to save the data for
3071        checking that an empty string has been matched. Previously this was taken
3072        from the 1000-entry workspace that had been reserved. So now there is no
3073        explicit limit, but more stack is used.
3074    
3075     8. Applied Daniel's patches to solve problems with the import/export magic
3076        syntax that is required for Windows, and which was going wrong for the
3077        pcreposix and pcrecpp parts of the library. These were overlooked when this
3078        problem was solved for the main library.
3079    
3080     9. There were some crude static tests to avoid integer overflow when computing
3081        the size of patterns that contain repeated groups with explicit upper
3082        limits. As the maximum quantifier is 65535, the maximum group length was
3083        set at 30,000 so that the product of these two numbers did not overflow a
3084        32-bit integer. However, it turns out that people want to use groups that
3085        are longer than 30,000 bytes (though not repeat them that many times).
3086        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3087        made it possible to implement the integer overflow checks in a much more
3088        dynamic way, which I have now done. The artificial limitation on group
3089        length has been removed - we now have only the limit on the total length of
3090        the compiled pattern, which depends on the LINK_SIZE setting.
3091    
3092    10. Fixed a bug in the documentation for get/copy named substring when
3093        duplicate names are permitted. If none of the named substrings are set, the
3094        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3095        empty string.
3096    
3097    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3098        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3099        because the ] is interpreted as the first data character and the
3100        terminating ] is not found. PCRE has been made compatible with Perl in this
3101        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3102        cause memory overwriting.
3103    
3104    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3105        string has been matched (to stop an infinite loop). It was not recognizing
3106        a conditional subpattern that could match an empty string if that
3107        subpattern was within another subpattern. For example, it looped when
3108        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3109        condition was not nested. This bug has been fixed.
3110    
3111    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3112        past the start of the subject in the presence of bytes with the top bit
3113        set, for example "\x8aBCD".
3114    
3115    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3116        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3117    
3118    14. Optimized (?!) to (*FAIL).
3119    
3120    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3121        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3122        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3123        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3124        does: it's just the validity check that is more restrictive.
3125    
3126    16. Inserted checks for integer overflows during escape sequence (backslash)
3127        processing, and also fixed erroneous offset values for syntax errors during
3128        backslash processing.
3129    
3130    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3131        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3132    
3133    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3134        caused an overrun.
3135    
3136    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3137        something other than just ASCII characters) inside a group that had an
3138        unlimited repeat caused a loop at compile time (while checking to see
3139        whether the group could match an empty string).
3140    
3141    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3142        [\P{Any}] did so. (Error in the code for printing property names.)
3143    
3144    21. An orphan \E inside a character class could cause a crash.
3145    
3146    22. A repeated capturing bracket such as (A)? could cause a wild memory
3147        reference during compilation.
3148    
3149    23. There are several functions in pcre_compile() that scan along a compiled
3150        expression for various reasons (e.g. to see if it's fixed length for look
3151        behind). There were bugs in these functions when a repeated \p or \P was
3152        present in the pattern. These operators have additional parameters compared
3153        with \d, etc, and these were not being taken into account when moving along
3154        the compiled data. Specifically:
3155    
3156        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3157            length.
3158    
3159        (b) An item such as \pL+ within a repeated group could cause crashes or
3160            loops.
3161    
3162        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3163            "reference to non-existent subpattern" error.
3164    
3165        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3166    
3167    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3168        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3169    
3170    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3171    
3172    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3173        character were causing crashes (broken optimization).
3174    
3175    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3176        \p or \P) caused a compile-time loop.
3177    
3178    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3179        break. For example, the unstudied pattern [\r\n]A does not match the string
3180        "\r\nA" because change 7.0/46 below moves the current point on by two
3181        characters after failing to match at the start. However, the pattern \nA
3182        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3183        the same is true. There doesn't seem any very clean way out of this, but
3184        what I have chosen to do makes the common cases work: PCRE now takes note
3185        of whether there can be an explicit match for \r or \n anywhere in the
3186        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3187        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3188        pattern has explicit CR or LF references.
3189    
3190    29. Added (*CR) etc for changing newline setting at start of pattern.
3191    
3192    
3193    Version 7.2 19-Jun-07
3194  ---------------------  ---------------------
3195    
3196   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 3210  Version 7.2 01-May-07
3210      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3211      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
3212      in all cases.      in all cases.
3213    
3214   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3215    
3216      (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
3217          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3218    
3219      (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
3220          to be opened parentheses.          to be opened parentheses.
3221    
3222      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3223          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3224    
3225        (d) \K resets the start of the current match so that everything before
3226            is not part of it.
3227    
3228        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3229    
3230        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3231            reference syntax.
3232    
3233        (g) (?| introduces a group in which the numbering of parentheses in each
3234            alternative starts with the same number.
3235    
3236        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3237    
3238     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3239        PCRE_INFO_JCHANGED.
3240    
3241     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3242        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3243        for detecting groups that can match an empty string.
3244    
3245     9. A pattern with a very large number of alternatives (more than several
3246        hundred) was running out of internal workspace during the pre-compile
3247        phase, where pcre_compile() figures out how much memory will be needed. A
3248        bit of new cunning has reduced the workspace needed for groups with
3249        alternatives. The 1000-alternative test pattern now uses 12 bytes of
3250        workspace instead of running out of the 4096 that are available.
3251    
3252    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3253    
3254    11. Applied patch from Google to remove an optimization that didn't quite work.
3255        The report of the bug said:
3256    
3257          pcrecpp::RE("a*").FullMatch("aaa") matches, while
3258          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3259          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3260    
3261    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3262        it matched the wrong number of bytes.
3263    
3264    
3265  Version 7.1 24-Apr-07  Version 7.1 24-Apr-07
3266  ---------------------  ---------------------

Legend:
Removed from v.167  
changed lines
  Added in v.1624

  ViewVC Help
Powered by ViewVC 1.1.5