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

Diff of /code/trunk/ChangeLog

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

revision 740 by zherczeg, Mon Oct 31 06:10:14 2011 UTC revision 1620 by ph10, Tue Dec 8 11:06:40 2015 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.21  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    
48    Version 8.38 23-November-2015
49    -----------------------------
50    
51    1.  If a group that contained a recursive back reference also contained a
52        forward reference subroutine call followed by a non-forward-reference
53        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
54        compile correct code, leading to undefined behaviour or an internally
55        detected error. This bug was discovered by the LLVM fuzzer.
56    
57    2.  Quantification of certain items (e.g. atomic back references) could cause
58        incorrect code to be compiled when recursive forward references were
59        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
60        This bug was discovered by the LLVM fuzzer.
61    
62    3.  A repeated conditional group whose condition was a reference by name caused
63        a buffer overflow if there was more than one group with the given name.
64        This bug was discovered by the LLVM fuzzer.
65    
66    4.  A recursive back reference by name within a group that had the same name as
67        another group caused a buffer overflow. For example:
68        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
69    
70    5.  A forward reference by name to a group whose number is the same as the
71        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
72        a buffer overflow at compile time. This bug was discovered by the LLVM
73        fuzzer.
74    
75    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
76        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
77    
78    7.  Another buffer overflow bug involved duplicate named groups with a
79        reference between their definition, with a group that reset capture
80        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
81        fixed by always allowing for more memory, even if not needed. (A proper fix
82        is implemented in PCRE2, but it involves more refactoring.)
83    
84    8.  There was no check for integer overflow in subroutine calls such as (?123).
85    
86    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
87        being treated as a literal 'l' instead of causing an error.
88    
89    10. There was a buffer overflow if pcre_exec() was called with an ovector of
90        size 1. This bug was found by american fuzzy lop.
91    
92    11. If a non-capturing group containing a conditional group that could match
93        an empty string was repeated, it was not identified as matching an empty
94        string itself. For example: /^(?:(?(1)x|)+)+$()/.
95    
96    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
97        \a and \e in test subject lines.
98    
99    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
100        instead of the EBCDIC value.
101    
102    14. The handling of \c in an EBCDIC environment has been revised so that it is
103        now compatible with the specification in Perl's perlebcdic page.
104    
105    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
106        ASCII/Unicode. This has now been added to the list of characters that are
107        recognized as white space in EBCDIC.
108    
109    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
110        error (correctly) when used outside a class, but did not give an error
111        within a class.
112    
113    17. \h within a class was incorrectly compiled in EBCDIC environments.
114    
115    18. A pattern with an unmatched closing parenthesis that contained a backward
116        assertion which itself contained a forward reference caused buffer
117        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
118    
119    19. JIT should return with error when the compiled pattern requires more stack
120        space than the maximum.
121    
122    20. A possessively repeated conditional group that could match an empty string,
123        for example, /(?(R))*+/, was incorrectly compiled.
124    
125    21. Fix infinite recursion in the JIT compiler when certain patterns such as
126        /(?:|a|){100}x/ are analysed.
127    
128    22. Some patterns with character classes involving [: and \\ were incorrectly
129        compiled and could cause reading from uninitialized memory or an incorrect
130        error diagnosis.
131    
132    23. Pathological patterns containing many nested occurrences of [: caused
133        pcre_compile() to run for a very long time.
134    
135    24. A conditional group with only one branch has an implicit empty alternative
136        branch and must therefore be treated as potentially matching an empty
137        string.
138    
139    25. If (?R was followed by - or + incorrect behaviour happened instead of a
140        diagnostic.
141    
142    26. Arrange to give up on finding the minimum matching length for overly
143        complex patterns.
144    
145    27. Similar to (4) above: in a pattern with duplicated named groups and an
146        occurrence of (?| it is possible for an apparently non-recursive back
147        reference to become recursive if a later named group with the relevant
148        number is encountered. This could lead to a buffer overflow. Wen Guanxing
149        from Venustech ADLAB discovered this bug.
150    
151    28. If pcregrep was given the -q option with -c or -l, or when handling a
152        binary file, it incorrectly wrote output to stdout.
153    
154    29. The JIT compiler did not restore the control verb head in case of *THEN
155        control verbs. This issue was found by Karl Skomski with a custom LLVM
156        fuzzer.
157    
158    30. Error messages for syntax errors following \g and \k were giving inaccurate
159        offsets in the pattern.
160    
161    31. Added a check for integer overflow in conditions (?(<digits>) and
162        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
163        fuzzer.
164    
165    32. Handling recursive references such as (?2) when the reference is to a group
166        later in the pattern uses code that is very hacked about and error-prone.
167        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
168        give an internal error if it is obvious that compiling has gone wrong.
169    
170    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
171        This issue was found by Karl Skomski with a custom LLVM fuzzer.
172    
173    34. The JIT compiler should restore the control chain for empty possessive
174        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
175    
176    35. Match limit check added to JIT recursion. This issue was found by Karl
177        Skomski with a custom LLVM fuzzer.
178    
179    36. Yet another case similar to 27 above has been circumvented by an
180        unconditional allocation of extra memory. This issue is fixed "properly" in
181        PCRE2 by refactoring the way references are handled. Wen Guanxing
182        from Venustech ADLAB discovered this bug.
183    
184    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
185        with a custom LLVM fuzzer.
186    
187    38. Fixed a corner case of range optimization in JIT.
188    
189    39. An incorrect error "overran compiling workspace" was given if there were
190        exactly enough group forward references such that the last one extended
191        into the workspace safety margin. The next one would have expanded the
192        workspace. The test for overflow was not including the safety margin.
193    
194    40. A match limit issue is fixed in JIT which was found by Karl Skomski
195        with a custom LLVM fuzzer.
196    
197    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
198        work under Windows. (Why has it taken so long for anyone to notice?)
199    
200    42. In a character class such as [\W\p{Any}] where both a negative-type escape
201        ("not a word character") and a property escape were present, the property
202        escape was being ignored.
203    
204    43. Fix crash caused by very long (*MARK) or (*THEN) names.
205    
206    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
207        by a single ASCII character in a class item, was incorrectly compiled in
208        UCP mode. The POSIX class got lost, but only if the single character
209        followed it.
210    
211    45. [:punct:] in UCP mode was matching some characters in the range 128-255
212        that should not have been matched.
213    
214    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
215        class, all characters with code points greater than 255 are in the class.
216        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
217        such as \w are turned into Unicode properties), wide characters were not
218        correctly handled, and could fail to match.
219    
220    
221    Version 8.37 28-April-2015
222    --------------------------
223    
224    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
225        for those parentheses to be closed with whatever has been captured so far.
226        However, it was failing to mark any other groups between the hightest
227        capture so far and the currrent group as "unset". Thus, the ovector for
228        those groups contained whatever was previously there. An example is the
229        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
230    
231    2.  If an assertion condition was quantified with a minimum of zero (an odd
232        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
233    
234    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
235        unrecognized modifier, a crash could occur.
236    
237    4.  An attempt to do global matching in pcretest with a zero-length ovector
238        caused a crash.
239    
240    5.  Fixed a memory leak during matching that could occur for a subpattern
241        subroutine call (recursive or otherwise) if the number of captured groups
242        that had to be saved was greater than ten.
243    
244    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
245        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
246        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
247    
248    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
249        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
250    
251    8. When a pattern is compiled, it remembers the highest back reference so that
252       when matching, if the ovector is too small, extra memory can be obtained to
253       use instead. A conditional subpattern whose condition is a check on a
254       capture having happened, such as, for example in the pattern
255       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
256       setting the highest backreference number. This mattered only if pcre_exec()
257       was called with an ovector that was too small to hold the capture, and there
258       was no other kind of back reference (a situation which is probably quite
259       rare). The effect of the bug was that the condition was always treated as
260       FALSE when the capture could not be consulted, leading to a incorrect
261       behaviour by pcre_exec(). This bug has been fixed.
262    
263    9. A reference to a duplicated named group (either a back reference or a test
264       for being set in a conditional) that occurred in a part of the pattern where
265       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
266       to be incorrectly calculated, leading to overwriting.
267    
268    10. A mutually recursive set of back references such as (\2)(\1) caused a
269        segfault at study time (while trying to find the minimum matching length).
270        The infinite loop is now broken (with the minimum length unset, that is,
271        zero).
272    
273    11. If an assertion that was used as a condition was quantified with a minimum
274        of zero, matching went wrong. In particular, if the whole group had
275        unlimited repetition and could match an empty string, a segfault was
276        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
277        assertions to be quantified, but not if they are being used as conditions,
278        so the above pattern is faulted by Perl. PCRE has now been changed so that
279        it also rejects such patterns.
280    
281    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
282        failed to allow the zero-repeat case if pcre2_exec() was called with an
283        ovector too small to capture the group.
284    
285    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
286        Red Hat Product Security:
287    
288        (a) A crash if /K and /F were both set with the option to save the compiled
289        pattern.
290    
291        (b) Another crash if the option to print captured substrings in a callout
292        was combined with setting a null ovector, for example \O\C+ as a subject
293        string.
294    
295    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
296        forward reference repeated a large (but limited) number of times within a
297        repeated outer group that has a zero minimum quantifier, caused incorrect
298        code to be compiled, leading to the error "internal error:
299        previously-checked referenced subpattern not found" when an incorrect
300        memory address was read. This bug was reported as "heap overflow",
301        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
302        CVE-2015-2325.
303    
304    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
305        call within a group that also contained a recursive back reference caused
306        incorrect code to be compiled. This bug was reported as "heap overflow",
307        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
308        number CVE-2015-2326.
309    
310    24. Computing the size of the JIT read-only data in advance has been a source
311        of various issues, and new ones are still appear unfortunately. To fix
312        existing and future issues, size computation is eliminated from the code,
313        and replaced by on-demand memory allocation.
314    
315    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
316        adjacent to the end of the range, and the range contained characters with
317        more than one other case, caused incorrect behaviour when compiled in UTF
318        mode. In that example, the range a-j was left out of the class.
319    
320    26. Fix JIT compilation of conditional blocks, which assertion
321        is converted to (*FAIL). E.g: /(?(?!))/.
322    
323    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
324        discovered by the LLVM fuzzer.
325    
326    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
327        when this assertion was used as a condition, for example (?(?!)a|b). In
328        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
329        error about an unsupported item.
330    
331    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
332        possessification code could take exponential time to complete. A recursion
333        depth limit of 1000 has been imposed to limit the resources used by this
334        optimization.
335    
336    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
337        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
338        because \S ensures they are all in the class. The code for doing this was
339        interacting badly with the code for computing the amount of space needed to
340        compile the pattern, leading to a buffer overflow. This bug was discovered
341        by the LLVM fuzzer.
342    
343    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
344        other kinds of group caused stack overflow at compile time. This bug was
345        discovered by the LLVM fuzzer.
346    
347    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
348        between a subroutine call and its quantifier was incorrectly compiled,
349        leading to buffer overflow or other errors. This bug was discovered by the
350        LLVM fuzzer.
351    
352    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
353        assertion after (?(. The code was failing to check the character after
354        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
355        was discovered by the LLVM fuzzer.
356    
357    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
358        a fixed maximum following a group that contains a subroutine reference was
359        incorrectly compiled and could trigger buffer overflow. This bug was
360        discovered by the LLVM fuzzer.
361    
362    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
363        caused a stack overflow instead of the diagnosis of a non-fixed length
364        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
365    
366    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
367        (e.g. /(?<=\Ka)/) could make pcregrep loop.
368    
369    37. There was a similar problem to 36 in pcretest for global matches.
370    
371    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
372        and a subsequent item in the pattern caused a non-match, backtracking over
373        the repeated \X did not stop, but carried on past the start of the subject,
374        causing reference to random memory and/or a segfault. There were also some
375        other cases where backtracking after \C could crash. This set of bugs was
376        discovered by the LLVM fuzzer.
377    
378    39. The function for finding the minimum length of a matching string could take
379        a very long time if mutual recursion was present many times in a pattern,
380        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
381        method has been implemented. This infelicity was discovered by the LLVM
382        fuzzer.
383    
384    40. Static linking against the PCRE library using the pkg-config module was
385        failing on missing pthread symbols.
386    
387    
388    Version 8.36 26-September-2014
389    ------------------------------
390    
391    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
392        -Wmissing-field-initializers and -Wunused-parameter.
393    
394    2.  The tests for quantifiers being too big (greater than 65535) were being
395        applied after reading the number, and stupidly assuming that integer
396        overflow would give a negative number. The tests are now applied as the
397        numbers are read.
398    
399    3.  Tidy code in pcre_exec.c where two branches that used to be different are
400        now the same.
401    
402    4.  The JIT compiler did not generate match limit checks for certain
403        bracketed expressions with quantifiers. This may lead to exponential
404        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
405        issue should be resolved now.
406    
407    5.  Fixed an issue, which occures when nested alternatives are optimized
408        with table jumps.
409    
410    6.  Inserted two casts and changed some ints to size_t in the light of some
411        reported 64-bit compiler warnings (Bugzilla 1477).
412    
413    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
414        an empty string, which sometimes were behaving incorrectly in the
415        interpreter (though correctly in the JIT matcher). This pcretest input is
416        an example:
417    
418          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
419          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
420    
421        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
422        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
423        for an empty string was breaking the inner loop and carrying on at a lower
424        level, when possessive repeated groups should always return to a higher
425        level as they have no backtrack points in them. The empty string test now
426        occurs at the outer level.
427    
428    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
429        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
430    
431    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
432        doesn't).
433    
434    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
435        to hold a pattern to be incorrectly computed (too small) when there were
436        named back references to duplicated names. This could cause "internal
437        error: code overflow" or "double free or corruption" or other memory
438        handling errors.
439    
440    11. When named subpatterns had the same prefixes, back references could be
441        confused. For example, in this pattern:
442    
443          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
444    
445        the reference to 'Name' was incorrectly treated as a reference to a
446        duplicate name.
447    
448    12. A pattern such as /^s?c/mi8 where the optional character has more than
449        one "other case" was incorrectly compiled such that it would only try to
450        match starting at "c".
451    
452    13. When a pattern starting with \s was studied, VT was not included in the
453        list of possible starting characters; this should have been part of the
454        8.34/18 patch.
455    
456    14. If a character class started [\Qx]... where x is any character, the class
457        was incorrectly terminated at the ].
458    
459    15. If a pattern that started with a caseless match for a character with more
460        than one "other case" was studied, PCRE did not set up the starting code
461        unit bit map for the list of possible characters. Now it does. This is an
462        optimization improvement, not a bug fix.
463    
464    16. The Unicode data tables have been updated to Unicode 7.0.0.
465    
466    17. Fixed a number of memory leaks in pcregrep.
467    
468    18. Avoid a compiler warning (from some compilers) for a function call with
469        a cast that removes "const" from an lvalue by using an intermediate
470        variable (to which the compiler does not object).
471    
472    19. Incorrect code was compiled if a group that contained an internal recursive
473        back reference was optional (had quantifier with a minimum of zero). This
474        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
475        caused segmentation faults because of stack overflows at compile time.
476    
477    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
478        group that is quantified with an indefinite repeat, caused a compile-time
479        loop which used up all the system stack and provoked a segmentation fault.
480        This was not the same bug as 19 above.
481    
482    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
483        Patch by Mike Frysinger.
484    
485    
486    Version 8.35 04-April-2014
487    --------------------------
488    
489    1.  A new flag is set, when property checks are present in an XCLASS.
490        When this flag is not set, PCRE can perform certain optimizations
491        such as studying these XCLASS-es.
492    
493    2.  The auto-possessification of character sets were improved: a normal
494        and an extended character set can be compared now. Furthermore
495        the JIT compiler optimizes more character set checks.
496    
497    3.  Got rid of some compiler warnings for potentially uninitialized variables
498        that show up only when compiled with -O2.
499    
500    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
501        of a match later then the end of the match. The pcretest program was not
502        handling the case sensibly - it was outputting from the start to the next
503        binary zero. It now reports this situation in a message, and outputs the
504        text from the end to the start.
505    
506    5.  Fast forward search is improved in JIT. Instead of the first three
507        characters, any three characters with fixed position can be searched.
508        Search order: first, last, middle.
509    
510    6.  Improve character range checks in JIT. Characters are read by an inprecise
511        function now, which returns with an unknown value if the character code is
512        above a certain threshold (e.g: 256). The only limitation is that the value
513        must be bigger than the threshold as well. This function is useful when
514        the characters above the threshold are handled in the same way.
515    
516    7.  The macros whose names start with RAWUCHAR are placeholders for a future
517        mode in which only the bottom 21 bits of 32-bit data items are used. To
518        make this more memorable for those maintaining the code, the names have
519        been changed to start with UCHAR21, and an extensive comment has been added
520        to their definition.
521    
522    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
523        to the export list in Makefile.am (they were accidentally omitted from the
524        8.34 tarball).
525    
526    9.  The informational output from pcretest used the phrase "starting byte set"
527        which is inappropriate for the 16-bit and 32-bit libraries. As the output
528        for "first char" and "need char" really means "non-UTF-char", I've changed
529        "byte" to "char", and slightly reworded the output. The documentation about
530        these values has also been (I hope) clarified.
531    
532    10. Another JIT related optimization: use table jumps for selecting the correct
533        backtracking path, when more than four alternatives are present inside a
534        bracket.
535    
536    11. Empty match is not possible, when the minimum length is greater than zero,
537        and there is no \K in the pattern. JIT should avoid empty match checks in
538        such cases.
539    
540    12. In a caseless character class with UCP support, when a character with more
541        than one alternative case was not the first character of a range, not all
542        the alternative cases were added to the class. For example, s and \x{17f}
543        are both alternative cases for S: the class [RST] was handled correctly,
544        but [R-T] was not.
545    
546    13. The configure.ac file always checked for pthread support when JIT was
547        enabled. This is not used in Windows, so I have put this test inside a
548        check for the presence of windows.h (which was already tested for).
549    
550    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
551        The algorithm provides a way to skip certain starting offsets, and usually
552        faster than linear prefix searches.
553    
554    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
555        as for 'fr_FR' and 'french'. For some reason, however, it then used the
556        Windows-specific input and output files, which have 'french' screwed in.
557        So this could never have worked. One of the problems with locales is that
558        they aren't always the same. I have now updated RunTest so that it checks
559        the output of the locale test (test 3) against three different output
560        files, and it allows the test to pass if any one of them matches. With luck
561        this should make the test pass on some versions of Solaris where it was
562        failing. Because of the uncertainty, the script did not used to stop if
563        test 3 failed; it now does. If further versions of a French locale ever
564        come to light, they can now easily be added.
565    
566    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
567        there was a message during ./configure, but it did not stop. This now
568        provokes an error. The invalid example in README has been corrected.
569        If a value less than the minimum is given, the minimum value has always
570        been used, but now a warning is given.
571    
572    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
573        was a bug in the test system, which is now fixed. Also, the list of various
574        configurations that are tested for each release did not have one with both
575        16/32 bits and --enable-bar-anycrlf. It now does.
576    
577    18. pcretest was missing "-C bsr" for displaying the \R default setting.
578    
579    19. Little endian PowerPC systems are supported now by the JIT compiler.
580    
581    20. The fast forward newline mechanism could enter to an infinite loop on
582        certain invalid UTF-8 input. Although we don't support these cases
583        this issue can be fixed by a performance optimization.
584    
585    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
586        not take account if existing stack usage. There is now a new global
587        variable called pcre_stack_guard that can be set to point to an external
588        function to check stack availability. It is called at the start of
589        processing every parenthesized group.
590    
591    22. A typo in the code meant that in ungreedy mode the max/min qualifier
592        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
593        not match "ab".
594    
595    23. When UTF was disabled, the JIT program reported some incorrect compile
596        errors. These messages are silenced now.
597    
598    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
599        compiler.
600    
601    25. Change all the temporary files used in RunGrepTest to be different to those
602        used by RunTest so that the tests can be run simultaneously, for example by
603        "make -j check".
604    
605    
606    Version 8.34 15-December-2013
607    -----------------------------
608    
609    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
610        executable memory. Patch inspired by Carsten Klein.
611    
612    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
613        this macro is never tested and has no effect, because the work to support
614        coverage involves only compiling and linking options and special targets in
615        the Makefile. The comment in config.h implied that defining the macro would
616        enable coverage support, which is totally false. There was also support for
617        setting this macro in the CMake files (my fault, I just copied it from
618        configure). SUPPORT_GCOV has now been removed.
619    
620    3.  Make a small performance improvement in strlen16() and strlen32() in
621        pcretest.
622    
623    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
624        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
625        these cases). There was also one in pcretest.c.
626    
627    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
628    
629    6.  In UTF mode, the code for checking whether a group could match an empty
630        string (which is used for indefinitely repeated groups to allow for
631        breaking an infinite loop) was broken when the group contained a repeated
632        negated single-character class with a character that occupied more than one
633        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
634        UTF-8 mode). The effect was undefined: the group might or might not be
635        deemed as matching an empty string, or the program might have crashed.
636    
637    7.  The code for checking whether a group could match an empty string was not
638        recognizing that \h, \H, \v, \V, and \R must match a character.
639    
640    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
641        an empty string. If it can, pcretest shows this in its information output.
642    
643    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
644        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
645        matched by pcre_exec() without using JIT:
646    
647        (a) If the rest of the pattern did not match after a maximal run of
648            grapheme clusters, the code for backing up to try with fewer of them
649            did not always back up over a full grapheme when characters that do not
650            have the modifier quality were involved, e.g. Hangul syllables.
651    
652        (b) If the match point in a subject started with modifier character, and
653            there was no match, the code could incorrectly back up beyond the match
654            point, and potentially beyond the first character in the subject,
655            leading to a segfault or an incorrect match result.
656    
657    10. A conditional group with an assertion condition could lead to PCRE
658        recording an incorrect first data item for a match if no other first data
659        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
660        first data item, and therefore matched "ca" after "c" instead of at the
661        start.
662    
663    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
664        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
665    
666    12. The source of pcregrep now includes z/OS-specific code so that it can be
667        compiled for z/OS as part of the special z/OS distribution.
668    
669    13. Added the -T and -TM options to pcretest.
670    
671    14. The code in pcre_compile.c for creating the table of named capturing groups
672        has been refactored. Instead of creating the table dynamically during the
673        actual compiling pass, the information is remembered during the pre-compile
674        pass (on the stack unless there are more than 20 named groups, in which
675        case malloc() is used) and the whole table is created before the actual
676        compile happens. This has simplified the code (it is now nearly 150 lines
677        shorter) and prepared the way for better handling of references to groups
678        with duplicate names.
679    
680    15. A back reference to a named subpattern when there is more than one of the
681        same name now checks them in the order in which they appear in the pattern.
682        The first one that is set is used for the reference. Previously only the
683        first one was inspected. This change makes PCRE more compatible with Perl.
684    
685    16. Unicode character properties were updated from Unicode 6.3.0.
686    
687    17. The compile-time code for auto-possessification has been refactored, based
688        on a patch by Zoltan Herczeg. It now happens after instead of during
689        compilation. The code is cleaner, and more cases are handled. The option
690        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
691        options in pcretest are provided to set it. It can also be set by
692        (*NO_AUTO_POSSESS) at the start of a pattern.
693    
694    18. The character VT has been added to the default ("C" locale) set of
695        characters that match \s and are generally treated as white space,
696        following this same change in Perl 5.18. There is now no difference between
697        "Perl space" and "POSIX space". Whether VT is treated as white space in
698        other locales depends on the locale.
699    
700    19. The code for checking named groups as conditions, either for being set or
701        for being recursed, has been refactored (this is related to 14 and 15
702        above). Processing unduplicated named groups should now be as fast at
703        numerical groups, and processing duplicated groups should be faster than
704        before.
705    
706    20. Two patches to the CMake build system, by Alexander Barkov:
707    
708          (1) Replace the "source" command by "." in CMakeLists.txt because
709              "source" is a bash-ism.
710    
711          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
712              without these the CMake build does not work on Solaris.
713    
714    21. Perl has changed its handling of \8 and \9. If there is no previously
715        encountered capturing group of those numbers, they are treated as the
716        literal characters 8 and 9 instead of a binary zero followed by the
717        literals. PCRE now does the same.
718    
719    22. Following Perl, added \o{} to specify codepoints in octal, making it
720        possible to specify values greater than 0777 and also making them
721        unambiguous.
722    
723    23. Perl now gives an error for missing closing braces after \x{... instead of
724        treating the string as literal. PCRE now does the same.
725    
726    24. RunTest used to grumble if an inappropriate test was selected explicitly,
727        but just skip it when running all tests. This make it awkward to run ranges
728        of tests when one of them was inappropriate. Now it just skips any
729        inappropriate tests, as it always did when running all tests.
730    
731    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
732        character types such as \d or \w, too many callouts were inserted, and the
733        data that they returned was rubbish.
734    
735    26. In UCP mode, \s was not matching two of the characters that Perl matches,
736        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
737        were matched by \h. The code has now been refactored so that the lists of
738        the horizontal and vertical whitespace characters used for \h and \v (which
739        are defined only in one place) are now also used for \s.
740    
741    27. Add JIT support for the 64 bit TileGX architecture.
742        Patch by Jiong Wang (Tilera Corporation).
743    
744    28. Possessive quantifiers for classes (both explicit and automatically
745        generated) now use special opcodes instead of wrapping in ONCE brackets.
746    
747    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
748        (because it's meaningless), this was not happening when PCRE_CASELESS was
749        set. Not wrong, but inefficient.
750    
751    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
752        properties for \w, \d, etc) is present in a test regex. Otherwise if the
753        test contains no characters greater than 255, Perl doesn't realise it
754        should be using Unicode semantics.
755    
756    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
757        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
758        does in Unicode mode.
759    
760    32. Added the "forbid" facility to pcretest so that putting tests into the
761        wrong test files can sometimes be quickly detected.
762    
763    33. There is now a limit (default 250) on the depth of nesting of parentheses.
764        This limit is imposed to control the amount of system stack used at compile
765        time. It can be changed at build time by --with-parens-nest-limit=xxx or
766        the equivalent in CMake.
767    
768    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
769        errors. Perl warns for these when in warning mode, but PCRE has no facility
770        for giving warnings.
771    
772    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
773        However, this was not working for (?!) because it is optimized to (*FAIL),
774        for which PCRE does not allow quantifiers. The optimization is now disabled
775        when a quantifier follows (?!). I can't see any use for this, but it makes
776        things uniform.
777    
778    36. Perl no longer allows group names to start with digits, so I have made this
779        change also in PCRE. It simplifies the code a bit.
780    
781    37. In extended mode, Perl ignores spaces before a + that indicates a
782        possessive quantifier. PCRE allowed a space before the quantifier, but not
783        before the possessive +. It now does.
784    
785    38. The use of \K (reset reported match start) within a repeated possessive
786        group such as (a\Kb)*+ was not working.
787    
788    40. Document that the same character tables must be used at compile time and
789        run time, and that the facility to pass tables to pcre_exec() and
790        pcre_dfa_exec() is for use only with saved/restored patterns.
791    
792    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
793        features for Builds with MSVC:
794    
795        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
796           builds). The .rc files can be used to set FileDescription and many other
797           attributes.
798    
799        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
800           This allows higher-level build scripts which want .pdb files to avoid
801           hard-coding the exact files needed."
802    
803    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
804        mean "start of word" and "end of word", respectively, as a transition aid.
805    
806    43. A minimizing repeat of a class containing codepoints greater than 255 in
807        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
808        compiled to use the heap for recursion.
809    
810    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
811        is configured.
812    
813    
814    Version 8.33 28-May-2013
815    ------------------------
816    
817    1.  Added 'U' to some constants that are compared to unsigned integers, to
818        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
819        variables that are added to signed variables, to ensure the result is
820        signed and can be negated.
821    
822    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
823        CMake config files.
824    
825    3.  Revise the creation of config.h.generic so that all boolean macros are
826        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
827        overriding via -D on the command line possible.
828    
829    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
830        to unsigned int is reported to make a quite noticeable speed difference in
831        a specific Windows environment. Testing on Linux did also appear to show
832        some benefit (and it is clearly not harmful). Also fixed the definition of
833        Xop which should be unsigned.
834    
835    5.  Related to (4), changing the definition of the intermediate variable cc
836        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
837        improvements.
838    
839    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
840        unnecessary spaces.
841    
842    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
843        and later.
844    
845    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
846    
847    9.  Optimizing fast_forward_start_bits in JIT.
848    
849    10. Adding support for callouts in JIT, and fixing some issues revealed
850        during this work. Namely:
851    
852        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
853    
854        (b) Minimum length was not checked before the matching is started.
855    
856    11. The value of capture_last that is passed to callouts was incorrect in some
857        cases when there was a capture on one path that was subsequently abandoned
858        after a backtrack. Also, the capture_last value is now reset after a
859        recursion, since all captures are also reset in this case.
860    
861    12. The interpreter no longer returns the "too many substrings" error in the
862        case when an overflowing capture is in a branch that is subsequently
863        abandoned after a backtrack.
864    
865    13. In the pathological case when an offset vector of size 2 is used, pcretest
866        now prints out the matched string after a yield of 0 or 1.
867    
868    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
869        Only supported by the JIT compiler at the moment.
870    
871    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
872    
873    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
874        offset of the starting point of the matching process, provided the offsets
875        vector is large enough.
876    
877    17. The \A escape now records a lookbehind value of 1, though its execution
878        does not actually inspect the previous character. This is to ensure that,
879        in partial multi-segment matching, at least one character from the old
880        segment is retained when a new segment is processed. Otherwise, if there
881        are no lookbehinds in the pattern, \A might match incorrectly at the start
882        of a new segment.
883    
884    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
885    
886    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
887        gives some modest performance improvement in 8-bit mode.
888    
889    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
890        be expressed in certain programming languages using Universal Character
891        Names.
892    
893    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
894        which points out that "non characters" are not "characters that may not
895        appear in Unicode strings" but rather "characters that are reserved for
896        internal use and have only local meaning".
897    
898    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
899        there was a conditional group that depended on an assertion, if the
900        assertion was false, the callout that immediately followed the alternation
901        in the condition was skipped when pcre_exec() was used for matching.
902    
903    23. Allow an explicit callout to be inserted before an assertion that is the
904        condition for a conditional group, for compatibility with automatic
905        callouts, which always insert a callout at this point.
906    
907    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
908        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
909    
910    25. (*PRUNE) is now supported by the JIT compiler.
911    
912    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
913    
914    27. Fix the case where there are two or more SKIPs with arguments that may be
915        ignored.
916    
917    28. (*SKIP) is now supported by the JIT compiler.
918    
919    29. (*THEN) is now supported by the JIT compiler.
920    
921    30. Update RunTest with additional test selector options.
922    
923    31. The way PCRE handles backtracking verbs has been changed in two ways.
924    
925        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
926        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
927        backtracking. In some cases this makes it more Perl-compatible, but Perl's
928        rather obscure rules do not always do the same thing.
929    
930        (2) Previously, backtracking verbs were confined within assertions. This is
931        no longer the case for positive assertions, except for (*ACCEPT). Again,
932        this sometimes improves Perl compatibility, and sometimes does not.
933    
934    32. A number of tests that were in test 2 because Perl did things differently
935        have been moved to test 1, because either Perl or PCRE has changed, and
936        these tests are now compatible.
937    
938    32. Backtracking control verbs are now handled in the same way in JIT and
939        interpreter.
940    
941    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
942        contained a forward subroutine reference caused a compile error.
943    
944    34. Auto-detect and optimize limited repetitions in JIT.
945    
946    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
947        blocking (*UTF) etc.
948    
949    36. In the interpreter, maximizing pattern repetitions for characters and
950        character types now use tail recursion, which reduces stack usage.
951    
952    37. The value of the max lookbehind was not correctly preserved if a compiled
953        and saved regex was reloaded on a host of different endianness.
954    
955    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
956        of the compiled pattern block, expand the flags field from 16 to 32 bits
957        because it was almost full.
958    
959    39. Try madvise first before posix_madvise.
960    
961    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
962        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
963        original change locked out all matches of empty strings. This has been
964        changed so that one match of an empty string per line is recognized.
965        Subsequent searches on the same line (for colouring or for --only-matching,
966        for example) do not recognize empty strings.
967    
968    41. Applied a user patch to fix a number of spelling mistakes in comments.
969    
970    42. Data lines longer than 65536 caused pcretest to crash.
971    
972    43. Clarified the data type for length and startoffset arguments for pcre_exec
973        and pcre_dfa_exec in the function-specific man pages, where they were
974        explicitly stated to be in bytes, never having been updated. I also added
975        some clarification to the pcreapi man page.
976    
977    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
978        a segmentation fault.
979    
980    
981    Version 8.32 30-November-2012
982    -----------------------------
983    
984    1.  Improved JIT compiler optimizations for first character search and single
985        character iterators.
986    
987    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
988        Patch by Daniel Richard G.
989    
990    3.  Single character iterator optimizations in the JIT compiler.
991    
992    4.  Improved JIT compiler optimizations for character ranges.
993    
994    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
995        Reported by Giuseppe D'Angelo.
996    
997    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
998        of a line, was being set incorrectly in cases where .* appeared inside
999        atomic brackets at the start of a pattern, or where there was a subsequent
1000        *PRUNE or *SKIP.
1001    
1002    7.  Improved instruction cache flush for POWER/PowerPC.
1003        Patch by Daniel Richard G.
1004    
1005    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1006        grep:
1007    
1008        (a) There is now no limit to the number of patterns to be matched.
1009    
1010        (b) An error is given if a pattern is too long.
1011    
1012        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1013            are now supported.
1014    
1015        (d) --exclude-from and --include-from (multiple use) have been added.
1016    
1017        (e) Exclusions and inclusions now apply to all files and directories, not
1018            just to those obtained from scanning a directory recursively.
1019    
1020        (f) Multiple uses of -f and --file-list are now supported.
1021    
1022        (g) In a Windows environment, the default for -d has been changed from
1023            "read" (the GNU grep default) to "skip", because otherwise the presence
1024            of a directory in the file list provokes an error.
1025    
1026        (h) The documentation has been revised and clarified in places.
1027    
1028    9.  Improve the matching speed of capturing brackets.
1029    
1030    10. Changed the meaning of \X so that it now matches a Unicode extended
1031        grapheme cluster.
1032    
1033    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1034        out POSIX threads when JIT support is configured.
1035    
1036    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1037    
1038    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1039        structure could go wrong in environments where size_t is not the same size
1040        as int.
1041    
1042    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1043        set.
1044    
1045    15. The EBCDIC support had decayed; later updates to the code had included
1046        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1047        general tidy up of EBCDIC-related issues, and the documentation was also
1048        not quite right. There is now a test that can be run on ASCII systems to
1049        check some of the EBCDIC-related things (but is it not a full test).
1050    
1051    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1052        in a small tidy to the code.
1053    
1054    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1055    
1056    18. If the --only-matching (-o) option in pcregrep is specified multiple
1057        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1058        substrings matched by the 1st and 2nd capturing parentheses. A separating
1059        string can be specified by --om-separator (default empty).
1060    
1061    19. Improving the first n character searches.
1062    
1063    20. Turn case lists for horizontal and vertical white space into macros so that
1064        they are defined only once.
1065    
1066    21. This set of changes together give more compatible Unicode case-folding
1067        behaviour for characters that have more than one other case when UCP
1068        support is available.
1069    
1070        (a) The Unicode property table now has offsets into a new table of sets of
1071            three or more characters that are case-equivalent. The MultiStage2.py
1072            script that generates these tables (the pcre_ucd.c file) now scans
1073            CaseFolding.txt instead of UnicodeData.txt for character case
1074            information.
1075    
1076        (b) The code for adding characters or ranges of characters to a character
1077            class has been abstracted into a generalized function that also handles
1078            case-independence. In UTF-mode with UCP support, this uses the new data
1079            to handle characters with more than one other case.
1080    
1081        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1082            whose other case is greater than 256 are now correctly matched
1083            caselessly. Previously, the high codepoint matched the low one, but not
1084            vice versa.
1085    
1086        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1087            of the new class addition function, using character lists defined as
1088            macros alongside the case definitions of 20 above.
1089    
1090        (e) Caseless back references now work with characters that have more than
1091            one other case.
1092    
1093        (f) General caseless matching of characters with more than one other case
1094            is supported.
1095    
1096    22. Unicode character properties were updated from Unicode 6.2.0
1097    
1098    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1099    
1100    24. Add support for 32-bit character strings, and UTF-32
1101    
1102    25. Major JIT compiler update (code refactoring and bugfixing).
1103        Experimental Sparc 32 support is added.
1104    
1105    26. Applied a modified version of Daniel Richard G's patch to create
1106        pcre.h.generic and config.h.generic by "make" instead of in the
1107        PrepareRelease script.
1108    
1109    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1110        pcre_compile.c when checking for a zero character.
1111    
1112    28. Introducing a native interface for JIT. Through this interface, the compiled
1113        machine code can be directly executed. The purpose of this interface is to
1114        provide fast pattern matching, so several sanity checks are not performed.
1115        However, feature tests are still performed. The new interface provides
1116        1.4x speedup compared to the old one.
1117    
1118    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1119        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1120        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1121    
1122    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1123        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1124        the "old" RFC 2279). Instead, it ended up passing a negative length to
1125        pcre_exec().
1126    
1127    31. Add support for GCC's visibility feature to hide internal functions.
1128    
1129    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1130        "unknown -C option" after outputting 0 or 1.
1131    
1132    33. There is now support for generating a code coverage report for the test
1133        suite in environments where gcc is the compiler and lcov is installed. This
1134        is mainly for the benefit of the developers.
1135    
1136    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1137        unaddressable using valgrind annotations, allowing valgrind to detect
1138        invalid memory accesses. This is mainly for the benefit of the developers.
1139    
1140    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1141    
1142    26. Give configure error if --enable-cpp but no C++ compiler found.
1143    
1144    
1145    Version 8.31 06-July-2012
1146    -------------------------
1147    
1148    1.  Fixing a wrong JIT test case and some compiler warnings.
1149    
1150    2.  Removed a bashism from the RunTest script.
1151    
1152    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1153        to unsigned type, result still unsigned" that was given by an MS compiler
1154        on encountering the code "-sizeof(xxx)".
1155    
1156    4.  Partial matching support is added to the JIT compiler.
1157    
1158    5.  Fixed several bugs concerned with partial matching of items that consist
1159        of more than one character:
1160    
1161        (a) /^(..)\1/ did not partially match "aba" because checking references was
1162            done on an "all or nothing" basis. This also applied to repeated
1163            references.
1164    
1165        (b) \R did not give a hard partial match if \r was found at the end of the
1166            subject.
1167    
1168        (c) \X did not give a hard partial match after matching one or more
1169            characters at the end of the subject.
1170    
1171        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1172            a partial match for the string "\r".
1173    
1174        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1175            a partial match for a CR character at the end of the subject string.
1176    
1177    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1178        running pcretest, the text "(JIT)" added to the output whenever JIT is
1179        actually used to run the match.
1180    
1181    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1182        or /S+[+] with a digit between 1 and 7.
1183    
1184    8.  OP_NOT now supports any UTF character not just single-byte ones.
1185    
1186    9.  (*MARK) control verb is now supported by the JIT compiler.
1187    
1188    10. The command "./RunTest list" lists the available tests without actually
1189        running any of them. (Because I keep forgetting what they all are.)
1190    
1191    11. Add PCRE_INFO_MAXLOOKBEHIND.
1192    
1193    12. Applied a (slightly modified) user-supplied patch that improves performance
1194        when the heap is used for recursion (compiled with --disable-stack-for-
1195        recursion). Instead of malloc and free for each heap frame each time a
1196        logical recursion happens, frames are retained on a chain and re-used where
1197        possible. This sometimes gives as much as 30% improvement.
1198    
1199    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1200        call.
1201    
1202    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1203    
1204    15. It is now possible to link pcretest with libedit as an alternative to
1205        libreadline.
1206    
1207    16. (*COMMIT) control verb is now supported by the JIT compiler.
1208    
1209    17. The Unicode data tables have been updated to Unicode 6.1.0.
1210    
1211    18. Added --file-list option to pcregrep.
1212    
1213    19. Added binary file support to pcregrep, including the -a, --binary-files,
1214        -I, and --text options.
1215    
1216    20. The madvise function is renamed for posix_madvise for QNX compatibility
1217        reasons. Fixed by Giuseppe D'Angelo.
1218    
1219    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1220    
1221    22. Changed the option for creating symbolic links for 16-bit man pages from
1222        -s to -sf so that re-installing does not cause issues.
1223    
1224    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1225    
1226    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1227        matches in certain environments (the workspace was not being correctly
1228        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1229        some of the workspace data at the beginning of a restart.
1230    
1231    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1232        was not doing so when it should - probably a typo introduced by SVN 528
1233        (change 8.10/14).
1234    
1235    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1236        \w+ when the character tables indicated that \x{c4} was a word character.
1237        There were several related cases, all because the tests for doing a table
1238        lookup were testing for characters less than 127 instead of 255.
1239    
1240    27. If a pattern contains capturing parentheses that are not used in a match,
1241        their slots in the ovector are set to -1. For those that are higher than
1242        any matched groups, this happens at the end of processing. In the case when
1243        there were back references that the ovector was too small to contain
1244        (causing temporary malloc'd memory to be used during matching), and the
1245        highest capturing number was not used, memory off the end of the ovector
1246        was incorrectly being set to -1. (It was using the size of the temporary
1247        memory instead of the true size.)
1248    
1249    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1250        ovector size, it uses memory at the end of the block that it has got.
1251    
1252    29. Check for an overlong MARK name and give an error at compile time. The
1253        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1254    
1255    30. JIT compiler update.
1256    
1257    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1258        Rill for the patch.
1259    
1260    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1261    
1262    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1263    
1264    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1265        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1266        was enabled.
1267    
1268    35. Improve JIT code generation for greedy plus quantifier.
1269    
1270    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1271        1 to "aa" instead of to an empty string. The bug affected repeated groups
1272        that could potentially match an empty string.
1273    
1274    37. Optimizing single character iterators in JIT.
1275    
1276    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1277        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1278        codepoints that are too big for the mode are faulted, and in a UTF mode,
1279        disallowed codepoints are also faulted.
1280    
1281    39. If PCRE was compiled with UTF support, in three places in the DFA
1282        matcher there was code that should only have been obeyed in UTF mode, but
1283        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1284        processing when bytes with values greater than 127 were present. In 16-bit
1285        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1286        both cases the values are those that cannot be the first data item in a UTF
1287        character. The three items that might have provoked this were recursions,
1288        possessively repeated groups, and atomic groups.
1289    
1290    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1291        and pcregrep, because some OS require shared objects to be explicitly
1292        passed to ld, causing the link step to fail if they are not.
1293    
1294    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1295        mode, patterns that started with \h* or \R* might be incorrectly matched.
1296    
1297    
1298    Version 8.30 04-February-2012
1299    -----------------------------
1300    
1301    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1302        name is defined in ctype.h.
1303    
1304    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1305        only in quite long subpatterns.
1306    
1307    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1308        since it was replaced by pcre_fullinfo() in February 2000.
1309    
1310    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1311        match a (*MARK), and the match failed at the start of the subject, a
1312        reference to memory before the start of the subject could occur. This bug
1313        was introduced by fix 17 of release 8.21.
1314    
1315    5.  A reference to an unset group with zero minimum repetition was giving
1316        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1317        /(another)?(\1?)test/ matched against "hello world test". This bug was
1318        introduced in release 8.13.
1319    
1320    6.  Add support for 16-bit character strings (a large amount of work involving
1321        many changes and refactorings).
1322    
1323    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1324        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1325        from a file.
1326    
1327    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1328        rounding is not applied in this particular case).
1329    
1330    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1331        if they appear, or are escaped, in patterns.
1332    
1333    10. Get rid of a number of -Wunused-but-set-variable warnings.
1334    
1335    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1336        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1337        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1338        also returns the mark "x". This bug applied to capturing parentheses,
1339        non-capturing parentheses, and atomic parentheses. It also applied to some
1340        assertions.
1341    
1342    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1343        information out of configure.ac instead of relying on pcre.h.generic, which
1344        is not stored in the repository.
1345    
1346    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1347        -lreadline.
1348    
1349    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1350    
1351    15. Applied Graycode's patch to put the top-level frame on the stack rather
1352        than the heap when not using the stack for recursion. This gives a
1353        performance improvement in many cases when recursion is not deep.
1354    
1355    16. Experimental code added to "pcretest -C" to output the stack frame size.
1356    
1357    
1358    Version 8.21 12-Dec-2011
1359    ------------------------
1360    
1361  1.  Updating the JIT compiler.  1.  Updating the JIT compiler.
1362    
1363    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1364        are added as well.
1365    
1366    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1367        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1368        calling _pcre_jit_exec. Some extra comments are added.
1369    
1370    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1371        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1372        was introduced by change 18 for 8.20.
1373    
1374    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1375        ECMA-262 standard.
1376    
1377    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1378        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1379        This bug was probably introduced by change 9 of 8.13.
1380    
1381    7.  While fixing 6 above, I noticed that a number of other items were being
1382        incorrectly rejected as "not fixed length". This arose partly because newer
1383        opcodes had not been added to the fixed-length checking code. I have (a)
1384        corrected the bug and added tests for these items, and (b) arranged for an
1385        error to occur if an unknown opcode is encountered while checking for fixed
1386        length instead of just assuming "not fixed length". The items that were
1387        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1388        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1389        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1390    
1391    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1392        being incorrectly compiled and would have given unpredicatble results.
1393    
1394    9.  A possessively repeated subpattern with minimum repeat count greater than
1395        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1396        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1397        the first (A) could occur when it should not.
1398    
1399    10. Add a cast and remove a redundant test from the code.
1400    
1401    11. JIT should use pcre_malloc/pcre_free for allocation.
1402    
1403    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1404        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1405        is anything other than /usr, -L is still shown).
1406    
1407    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1408    
1409    14. Perl does not support \N without a following name in a [] class; PCRE now
1410        also gives an error.
1411    
1412    15. If a forward reference was repeated with an upper limit of around 2000,
1413        it caused the error "internal error: overran compiling workspace". The
1414        maximum number of forward references (including repeats) was limited by the
1415        internal workspace, and dependent on the LINK_SIZE. The code has been
1416        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1417        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1418        of around 200,000 forward references. While doing this, I also speeded up
1419        the filling in of repeated forward references.
1420    
1421    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1422        incorrectly expecting the subject to contain another "a" after the start.
1423    
1424    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1425        in the match, the SKIP should be ignored. This was not happening; instead
1426        the SKIP was being treated as NOMATCH. For patterns such as
1427        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1428        tested.
1429    
1430    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1431        now much more compatible with Perl, in particular in cases where the result
1432        is a non-match for a non-anchored pattern. For example, if
1433        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1434        "m", where previously it did not return a name. A side effect of this
1435        change is that for partial matches, the last encountered mark name is
1436        returned, as for non matches. A number of tests that were previously not
1437        Perl-compatible have been moved into the Perl-compatible test files. The
1438        refactoring has had the pleasing side effect of removing one argument from
1439        the match() function, thus reducing its stack requirements.
1440    
1441    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1442        subsequent uses of /S (without +) incorrectly behaved like /S+.
1443    
1444    21. Retrieve executable code size support for the JIT compiler and fixing
1445        some warnings.
1446    
1447    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1448        not work when the shorter character appeared right at the end of the
1449        subject string.
1450    
1451    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1452        systems.
1453    
1454    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1455        output it when the /M option is used in pcretest.
1456    
1457    25. The CheckMan script was not being included in the distribution. Also, added
1458        an explicit "perl" to run Perl scripts from the PrepareRelease script
1459        because this is reportedly needed in Windows.
1460    
1461    26. If study data was being save in a file and studying had not found a set of
1462        "starts with" bytes for the pattern, the data written to the file (though
1463        never used) was taken from uninitialized memory and so caused valgrind to
1464        complain.
1465    
1466    27. Updated RunTest.bat as provided by Sheri Pierce.
1467    
1468    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1469    
1470    29. Computation of memory usage for the table of capturing group names was
1471        giving an unnecessarily large value.
1472    
1473    
1474  Version 8.20 21-Oct-2011  Version 8.20 21-Oct-2011
1475  ------------------------  ------------------------
# Line 1021  Version 7.9 11-Apr-09 Line 2485  Version 7.9 11-Apr-09
2485  7.  A pattern that could match an empty string could cause pcregrep to loop; it  7.  A pattern that could match an empty string could cause pcregrep to loop; it
2486      doesn't make sense to accept an empty string match in pcregrep, so I have      doesn't make sense to accept an empty string match in pcregrep, so I have
2487      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this      locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2488      seems to be how GNU grep behaves.      seems to be how GNU grep behaves. [But see later change 40 for release
2489        8.33.]
2490    
2491  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at  8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2492      start or after a newline", because the conditional assertion was not being      start or after a newline", because the conditional assertion was not being
# Line 1264  Version 7.7 07-May-08 Line 2729  Version 7.7 07-May-08
2729      containing () gave an internal compiling error instead of "reference to      containing () gave an internal compiling error instead of "reference to
2730      non-existent subpattern". Fortunately, when the pattern did exist, the      non-existent subpattern". Fortunately, when the pattern did exist, the
2731      compiled code was correct. (When scanning forwards to check for the      compiled code was correct. (When scanning forwards to check for the
2732      existencd of the subpattern, it was treating the data ']' as terminating      existence of the subpattern, it was treating the data ']' as terminating
2733      the class, so got the count wrong. When actually compiling, the reference      the class, so got the count wrong. When actually compiling, the reference
2734      was subsequently set up correctly.)      was subsequently set up correctly.)
2735    

Legend:
Removed from v.740  
changed lines
  Added in v.1620

  ViewVC Help
Powered by ViewVC 1.1.5