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

Diff of /code/trunk/ChangeLog

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

revision 1388 by ph10, Tue Nov 5 15:24:45 2013 UTC revision 1658 by ph10, Fri Jun 17 17:28:14 2016 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 8.34 xx-xxxx-201x  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 8.40 17-June-2016
8    -------------------------
9    
10    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
11        the match extended over a line boundary.
12    
13    
14    Version 8.39 14-June-2016
15    -------------------------
16    
17    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
18        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
19        misbehaved. This bug was found by the LLVM fuzzer.
20    
21    2.  Similar to the above, if an isolated \E was present between an item and its
22        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
23        bug was found by the LLVM fuzzer.
24    
25    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
26        working correctly in UCP mode.
27    
28    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
29        was set when the pmatch argument was NULL. It now returns REG_INVARG.
30    
31    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
32    
33    6.  An empty \Q\E sequence between an item and its qualifier caused
34        pcre_compile() to misbehave when auto callouts were enabled. This bug was
35        found by the LLVM fuzzer.
36    
37    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
38        space or a #-type comment that was followed by (?-x), which turns off
39        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
40        pcre_compile() assumed that (?-x) applied to the whole pattern and
41        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
42    
43    8.  A call of pcre_copy_named_substring() for a named substring whose number
44        was greater than the space in the ovector could cause a crash.
45    
46    9.  Yet another buffer overflow bug involved duplicate named groups with a
47        group that reset capture numbers (compare 8.38/7 below). Once again, I have
48        just allowed for more memory, even if not needed. (A proper fix is
49        implemented in PCRE2, but it involves a lot of refactoring.)
50    
51    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
52        start of the match to be earlier than the end.
53    
54    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
55    
56    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
57        assertion, caused pcretest to generate incorrect output, and also to read
58        uninitialized memory (detected by ASAN or valgrind).
59    
60    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
61        nested set of parentheses of sufficient size caused an overflow of the
62        compiling workspace (which was diagnosed, but of course is not desirable).
63    
64    14. And yet another buffer overflow bug involving duplicate named groups, this
65        time nested, with a nested back reference. Yet again, I have just allowed
66        for more memory, because anything more needs all the refactoring that has
67        been done for PCRE2. An example pattern that provoked this bug is:
68        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
69        registered as CVE-2016-1283.
70    
71    15. pcretest went into a loop if global matching was requested with an ovector
72        size less than 2. It now gives an error message. This bug was found by
73        afl-fuzz.
74    
75    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
76        ("assertion expected") when (?(?C) was not followed by an opening
77        parenthesis.
78    
79    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
80        actually affect anything, by sheer luck.
81    
82    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
83        static compilation.
84    
85    19. Modified the RunTest script to incorporate a valgrind suppressions file so
86        that certain errors, provoked by the SSE2 instruction set when JIT is used,
87        are ignored.
88    
89    20. A racing condition is fixed in JIT reported by Mozilla.
90    
91    21. Minor code refactor to avoid "array subscript is below array bounds"
92        compiler warning.
93    
94    22. Minor code refactor to avoid "left shift of negative number" warning.
95    
96    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
97        UCP support.
98    
99    24. Refactor to avoid compiler warnings in pcrecpp.cc.
100    
101    25. Refactor to fix a typo in pcre_jit_test.c
102    
103    26. Patch to support compiling pcrecpp.cc with Intel compiler.
104    
105    
106    Version 8.38 23-November-2015
107    -----------------------------
108    
109    1.  If a group that contained a recursive back reference also contained a
110        forward reference subroutine call followed by a non-forward-reference
111        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
112        compile correct code, leading to undefined behaviour or an internally
113        detected error. This bug was discovered by the LLVM fuzzer.
114    
115    2.  Quantification of certain items (e.g. atomic back references) could cause
116        incorrect code to be compiled when recursive forward references were
117        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
118        This bug was discovered by the LLVM fuzzer.
119    
120    3.  A repeated conditional group whose condition was a reference by name caused
121        a buffer overflow if there was more than one group with the given name.
122        This bug was discovered by the LLVM fuzzer.
123    
124    4.  A recursive back reference by name within a group that had the same name as
125        another group caused a buffer overflow. For example:
126        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
127    
128    5.  A forward reference by name to a group whose number is the same as the
129        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
130        a buffer overflow at compile time. This bug was discovered by the LLVM
131        fuzzer.
132    
133    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
134        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
135    
136    7.  Another buffer overflow bug involved duplicate named groups with a
137        reference between their definition, with a group that reset capture
138        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
139        fixed by always allowing for more memory, even if not needed. (A proper fix
140        is implemented in PCRE2, but it involves more refactoring.)
141    
142    8.  There was no check for integer overflow in subroutine calls such as (?123).
143    
144    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
145        being treated as a literal 'l' instead of causing an error.
146    
147    10. There was a buffer overflow if pcre_exec() was called with an ovector of
148        size 1. This bug was found by american fuzzy lop.
149    
150    11. If a non-capturing group containing a conditional group that could match
151        an empty string was repeated, it was not identified as matching an empty
152        string itself. For example: /^(?:(?(1)x|)+)+$()/.
153    
154    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
155        \a and \e in test subject lines.
156    
157    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
158        instead of the EBCDIC value.
159    
160    14. The handling of \c in an EBCDIC environment has been revised so that it is
161        now compatible with the specification in Perl's perlebcdic page.
162    
163    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
164        ASCII/Unicode. This has now been added to the list of characters that are
165        recognized as white space in EBCDIC.
166    
167    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
168        error (correctly) when used outside a class, but did not give an error
169        within a class.
170    
171    17. \h within a class was incorrectly compiled in EBCDIC environments.
172    
173    18. A pattern with an unmatched closing parenthesis that contained a backward
174        assertion which itself contained a forward reference caused buffer
175        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
176    
177    19. JIT should return with error when the compiled pattern requires more stack
178        space than the maximum.
179    
180    20. A possessively repeated conditional group that could match an empty string,
181        for example, /(?(R))*+/, was incorrectly compiled.
182    
183    21. Fix infinite recursion in the JIT compiler when certain patterns such as
184        /(?:|a|){100}x/ are analysed.
185    
186    22. Some patterns with character classes involving [: and \\ were incorrectly
187        compiled and could cause reading from uninitialized memory or an incorrect
188        error diagnosis.
189    
190    23. Pathological patterns containing many nested occurrences of [: caused
191        pcre_compile() to run for a very long time.
192    
193    24. A conditional group with only one branch has an implicit empty alternative
194        branch and must therefore be treated as potentially matching an empty
195        string.
196    
197    25. If (?R was followed by - or + incorrect behaviour happened instead of a
198        diagnostic.
199    
200    26. Arrange to give up on finding the minimum matching length for overly
201        complex patterns.
202    
203    27. Similar to (4) above: in a pattern with duplicated named groups and an
204        occurrence of (?| it is possible for an apparently non-recursive back
205        reference to become recursive if a later named group with the relevant
206        number is encountered. This could lead to a buffer overflow. Wen Guanxing
207        from Venustech ADLAB discovered this bug.
208    
209    28. If pcregrep was given the -q option with -c or -l, or when handling a
210        binary file, it incorrectly wrote output to stdout.
211    
212    29. The JIT compiler did not restore the control verb head in case of *THEN
213        control verbs. This issue was found by Karl Skomski with a custom LLVM
214        fuzzer.
215    
216    30. Error messages for syntax errors following \g and \k were giving inaccurate
217        offsets in the pattern.
218    
219    31. Added a check for integer overflow in conditions (?(<digits>) and
220        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
221        fuzzer.
222    
223    32. Handling recursive references such as (?2) when the reference is to a group
224        later in the pattern uses code that is very hacked about and error-prone.
225        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
226        give an internal error if it is obvious that compiling has gone wrong.
227    
228    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
229        This issue was found by Karl Skomski with a custom LLVM fuzzer.
230    
231    34. The JIT compiler should restore the control chain for empty possessive
232        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
233    
234    35. Match limit check added to JIT recursion. This issue was found by Karl
235        Skomski with a custom LLVM fuzzer.
236    
237    36. Yet another case similar to 27 above has been circumvented by an
238        unconditional allocation of extra memory. This issue is fixed "properly" in
239        PCRE2 by refactoring the way references are handled. Wen Guanxing
240        from Venustech ADLAB discovered this bug.
241    
242    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
243        with a custom LLVM fuzzer.
244    
245    38. Fixed a corner case of range optimization in JIT.
246    
247    39. An incorrect error "overran compiling workspace" was given if there were
248        exactly enough group forward references such that the last one extended
249        into the workspace safety margin. The next one would have expanded the
250        workspace. The test for overflow was not including the safety margin.
251    
252    40. A match limit issue is fixed in JIT which was found by Karl Skomski
253        with a custom LLVM fuzzer.
254    
255    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
256        work under Windows. (Why has it taken so long for anyone to notice?)
257    
258    42. In a character class such as [\W\p{Any}] where both a negative-type escape
259        ("not a word character") and a property escape were present, the property
260        escape was being ignored.
261    
262    43. Fix crash caused by very long (*MARK) or (*THEN) names.
263    
264    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
265        by a single ASCII character in a class item, was incorrectly compiled in
266        UCP mode. The POSIX class got lost, but only if the single character
267        followed it.
268    
269    45. [:punct:] in UCP mode was matching some characters in the range 128-255
270        that should not have been matched.
271    
272    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
273        class, all characters with code points greater than 255 are in the class.
274        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
275        such as \w are turned into Unicode properties), wide characters were not
276        correctly handled, and could fail to match.
277    
278    
279    Version 8.37 28-April-2015
280    --------------------------
281    
282    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
283        for those parentheses to be closed with whatever has been captured so far.
284        However, it was failing to mark any other groups between the hightest
285        capture so far and the currrent group as "unset". Thus, the ovector for
286        those groups contained whatever was previously there. An example is the
287        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
288    
289    2.  If an assertion condition was quantified with a minimum of zero (an odd
290        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
291    
292    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
293        unrecognized modifier, a crash could occur.
294    
295    4.  An attempt to do global matching in pcretest with a zero-length ovector
296        caused a crash.
297    
298    5.  Fixed a memory leak during matching that could occur for a subpattern
299        subroutine call (recursive or otherwise) if the number of captured groups
300        that had to be saved was greater than ten.
301    
302    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
303        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
304        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
305    
306    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
307        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
308    
309    8. When a pattern is compiled, it remembers the highest back reference so that
310       when matching, if the ovector is too small, extra memory can be obtained to
311       use instead. A conditional subpattern whose condition is a check on a
312       capture having happened, such as, for example in the pattern
313       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
314       setting the highest backreference number. This mattered only if pcre_exec()
315       was called with an ovector that was too small to hold the capture, and there
316       was no other kind of back reference (a situation which is probably quite
317       rare). The effect of the bug was that the condition was always treated as
318       FALSE when the capture could not be consulted, leading to a incorrect
319       behaviour by pcre_exec(). This bug has been fixed.
320    
321    9. A reference to a duplicated named group (either a back reference or a test
322       for being set in a conditional) that occurred in a part of the pattern where
323       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
324       to be incorrectly calculated, leading to overwriting.
325    
326    10. A mutually recursive set of back references such as (\2)(\1) caused a
327        segfault at study time (while trying to find the minimum matching length).
328        The infinite loop is now broken (with the minimum length unset, that is,
329        zero).
330    
331    11. If an assertion that was used as a condition was quantified with a minimum
332        of zero, matching went wrong. In particular, if the whole group had
333        unlimited repetition and could match an empty string, a segfault was
334        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
335        assertions to be quantified, but not if they are being used as conditions,
336        so the above pattern is faulted by Perl. PCRE has now been changed so that
337        it also rejects such patterns.
338    
339    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
340        failed to allow the zero-repeat case if pcre2_exec() was called with an
341        ovector too small to capture the group.
342    
343    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
344        Red Hat Product Security:
345    
346        (a) A crash if /K and /F were both set with the option to save the compiled
347        pattern.
348    
349        (b) Another crash if the option to print captured substrings in a callout
350        was combined with setting a null ovector, for example \O\C+ as a subject
351        string.
352    
353    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
354        forward reference repeated a large (but limited) number of times within a
355        repeated outer group that has a zero minimum quantifier, caused incorrect
356        code to be compiled, leading to the error "internal error:
357        previously-checked referenced subpattern not found" when an incorrect
358        memory address was read. This bug was reported as "heap overflow",
359        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
360        CVE-2015-2325.
361    
362    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
363        call within a group that also contained a recursive back reference caused
364        incorrect code to be compiled. This bug was reported as "heap overflow",
365        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
366        number CVE-2015-2326.
367    
368    24. Computing the size of the JIT read-only data in advance has been a source
369        of various issues, and new ones are still appear unfortunately. To fix
370        existing and future issues, size computation is eliminated from the code,
371        and replaced by on-demand memory allocation.
372    
373    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
374        adjacent to the end of the range, and the range contained characters with
375        more than one other case, caused incorrect behaviour when compiled in UTF
376        mode. In that example, the range a-j was left out of the class.
377    
378    26. Fix JIT compilation of conditional blocks, which assertion
379        is converted to (*FAIL). E.g: /(?(?!))/.
380    
381    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
382        discovered by the LLVM fuzzer.
383    
384    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
385        when this assertion was used as a condition, for example (?(?!)a|b). In
386        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
387        error about an unsupported item.
388    
389    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
390        possessification code could take exponential time to complete. A recursion
391        depth limit of 1000 has been imposed to limit the resources used by this
392        optimization.
393    
394    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
395        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
396        because \S ensures they are all in the class. The code for doing this was
397        interacting badly with the code for computing the amount of space needed to
398        compile the pattern, leading to a buffer overflow. This bug was discovered
399        by the LLVM fuzzer.
400    
401    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
402        other kinds of group caused stack overflow at compile time. This bug was
403        discovered by the LLVM fuzzer.
404    
405    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
406        between a subroutine call and its quantifier was incorrectly compiled,
407        leading to buffer overflow or other errors. This bug was discovered by the
408        LLVM fuzzer.
409    
410    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
411        assertion after (?(. The code was failing to check the character after
412        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
413        was discovered by the LLVM fuzzer.
414    
415    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
416        a fixed maximum following a group that contains a subroutine reference was
417        incorrectly compiled and could trigger buffer overflow. This bug was
418        discovered by the LLVM fuzzer.
419    
420    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
421        caused a stack overflow instead of the diagnosis of a non-fixed length
422        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
423    
424    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
425        (e.g. /(?<=\Ka)/) could make pcregrep loop.
426    
427    37. There was a similar problem to 36 in pcretest for global matches.
428    
429    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
430        and a subsequent item in the pattern caused a non-match, backtracking over
431        the repeated \X did not stop, but carried on past the start of the subject,
432        causing reference to random memory and/or a segfault. There were also some
433        other cases where backtracking after \C could crash. This set of bugs was
434        discovered by the LLVM fuzzer.
435    
436    39. The function for finding the minimum length of a matching string could take
437        a very long time if mutual recursion was present many times in a pattern,
438        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
439        method has been implemented. This infelicity was discovered by the LLVM
440        fuzzer.
441    
442    40. Static linking against the PCRE library using the pkg-config module was
443        failing on missing pthread symbols.
444    
445    
446    Version 8.36 26-September-2014
447    ------------------------------
448    
449    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
450        -Wmissing-field-initializers and -Wunused-parameter.
451    
452    2.  The tests for quantifiers being too big (greater than 65535) were being
453        applied after reading the number, and stupidly assuming that integer
454        overflow would give a negative number. The tests are now applied as the
455        numbers are read.
456    
457    3.  Tidy code in pcre_exec.c where two branches that used to be different are
458        now the same.
459    
460    4.  The JIT compiler did not generate match limit checks for certain
461        bracketed expressions with quantifiers. This may lead to exponential
462        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
463        issue should be resolved now.
464    
465    5.  Fixed an issue, which occures when nested alternatives are optimized
466        with table jumps.
467    
468    6.  Inserted two casts and changed some ints to size_t in the light of some
469        reported 64-bit compiler warnings (Bugzilla 1477).
470    
471    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
472        an empty string, which sometimes were behaving incorrectly in the
473        interpreter (though correctly in the JIT matcher). This pcretest input is
474        an example:
475    
476          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
477          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
478    
479        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
480        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
481        for an empty string was breaking the inner loop and carrying on at a lower
482        level, when possessive repeated groups should always return to a higher
483        level as they have no backtrack points in them. The empty string test now
484        occurs at the outer level.
485    
486    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
487        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
488    
489    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
490        doesn't).
491    
492    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
493        to hold a pattern to be incorrectly computed (too small) when there were
494        named back references to duplicated names. This could cause "internal
495        error: code overflow" or "double free or corruption" or other memory
496        handling errors.
497    
498    11. When named subpatterns had the same prefixes, back references could be
499        confused. For example, in this pattern:
500    
501          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
502    
503        the reference to 'Name' was incorrectly treated as a reference to a
504        duplicate name.
505    
506    12. A pattern such as /^s?c/mi8 where the optional character has more than
507        one "other case" was incorrectly compiled such that it would only try to
508        match starting at "c".
509    
510    13. When a pattern starting with \s was studied, VT was not included in the
511        list of possible starting characters; this should have been part of the
512        8.34/18 patch.
513    
514    14. If a character class started [\Qx]... where x is any character, the class
515        was incorrectly terminated at the ].
516    
517    15. If a pattern that started with a caseless match for a character with more
518        than one "other case" was studied, PCRE did not set up the starting code
519        unit bit map for the list of possible characters. Now it does. This is an
520        optimization improvement, not a bug fix.
521    
522    16. The Unicode data tables have been updated to Unicode 7.0.0.
523    
524    17. Fixed a number of memory leaks in pcregrep.
525    
526    18. Avoid a compiler warning (from some compilers) for a function call with
527        a cast that removes "const" from an lvalue by using an intermediate
528        variable (to which the compiler does not object).
529    
530    19. Incorrect code was compiled if a group that contained an internal recursive
531        back reference was optional (had quantifier with a minimum of zero). This
532        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
533        caused segmentation faults because of stack overflows at compile time.
534    
535    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
536        group that is quantified with an indefinite repeat, caused a compile-time
537        loop which used up all the system stack and provoked a segmentation fault.
538        This was not the same bug as 19 above.
539    
540    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
541        Patch by Mike Frysinger.
542    
543    
544    Version 8.35 04-April-2014
545  --------------------------  --------------------------
546    
547    1.  A new flag is set, when property checks are present in an XCLASS.
548        When this flag is not set, PCRE can perform certain optimizations
549        such as studying these XCLASS-es.
550    
551    2.  The auto-possessification of character sets were improved: a normal
552        and an extended character set can be compared now. Furthermore
553        the JIT compiler optimizes more character set checks.
554    
555    3.  Got rid of some compiler warnings for potentially uninitialized variables
556        that show up only when compiled with -O2.
557    
558    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
559        of a match later then the end of the match. The pcretest program was not
560        handling the case sensibly - it was outputting from the start to the next
561        binary zero. It now reports this situation in a message, and outputs the
562        text from the end to the start.
563    
564    5.  Fast forward search is improved in JIT. Instead of the first three
565        characters, any three characters with fixed position can be searched.
566        Search order: first, last, middle.
567    
568    6.  Improve character range checks in JIT. Characters are read by an inprecise
569        function now, which returns with an unknown value if the character code is
570        above a certain threshold (e.g: 256). The only limitation is that the value
571        must be bigger than the threshold as well. This function is useful when
572        the characters above the threshold are handled in the same way.
573    
574    7.  The macros whose names start with RAWUCHAR are placeholders for a future
575        mode in which only the bottom 21 bits of 32-bit data items are used. To
576        make this more memorable for those maintaining the code, the names have
577        been changed to start with UCHAR21, and an extensive comment has been added
578        to their definition.
579    
580    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
581        to the export list in Makefile.am (they were accidentally omitted from the
582        8.34 tarball).
583    
584    9.  The informational output from pcretest used the phrase "starting byte set"
585        which is inappropriate for the 16-bit and 32-bit libraries. As the output
586        for "first char" and "need char" really means "non-UTF-char", I've changed
587        "byte" to "char", and slightly reworded the output. The documentation about
588        these values has also been (I hope) clarified.
589    
590    10. Another JIT related optimization: use table jumps for selecting the correct
591        backtracking path, when more than four alternatives are present inside a
592        bracket.
593    
594    11. Empty match is not possible, when the minimum length is greater than zero,
595        and there is no \K in the pattern. JIT should avoid empty match checks in
596        such cases.
597    
598    12. In a caseless character class with UCP support, when a character with more
599        than one alternative case was not the first character of a range, not all
600        the alternative cases were added to the class. For example, s and \x{17f}
601        are both alternative cases for S: the class [RST] was handled correctly,
602        but [R-T] was not.
603    
604    13. The configure.ac file always checked for pthread support when JIT was
605        enabled. This is not used in Windows, so I have put this test inside a
606        check for the presence of windows.h (which was already tested for).
607    
608    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
609        The algorithm provides a way to skip certain starting offsets, and usually
610        faster than linear prefix searches.
611    
612    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
613        as for 'fr_FR' and 'french'. For some reason, however, it then used the
614        Windows-specific input and output files, which have 'french' screwed in.
615        So this could never have worked. One of the problems with locales is that
616        they aren't always the same. I have now updated RunTest so that it checks
617        the output of the locale test (test 3) against three different output
618        files, and it allows the test to pass if any one of them matches. With luck
619        this should make the test pass on some versions of Solaris where it was
620        failing. Because of the uncertainty, the script did not used to stop if
621        test 3 failed; it now does. If further versions of a French locale ever
622        come to light, they can now easily be added.
623    
624    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
625        there was a message during ./configure, but it did not stop. This now
626        provokes an error. The invalid example in README has been corrected.
627        If a value less than the minimum is given, the minimum value has always
628        been used, but now a warning is given.
629    
630    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
631        was a bug in the test system, which is now fixed. Also, the list of various
632        configurations that are tested for each release did not have one with both
633        16/32 bits and --enable-bar-anycrlf. It now does.
634    
635    18. pcretest was missing "-C bsr" for displaying the \R default setting.
636    
637    19. Little endian PowerPC systems are supported now by the JIT compiler.
638    
639    20. The fast forward newline mechanism could enter to an infinite loop on
640        certain invalid UTF-8 input. Although we don't support these cases
641        this issue can be fixed by a performance optimization.
642    
643    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
644        not take account if existing stack usage. There is now a new global
645        variable called pcre_stack_guard that can be set to point to an external
646        function to check stack availability. It is called at the start of
647        processing every parenthesized group.
648    
649    22. A typo in the code meant that in ungreedy mode the max/min qualifier
650        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
651        not match "ab".
652    
653    23. When UTF was disabled, the JIT program reported some incorrect compile
654        errors. These messages are silenced now.
655    
656    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
657        compiler.
658    
659    25. Change all the temporary files used in RunGrepTest to be different to those
660        used by RunTest so that the tests can be run simultaneously, for example by
661        "make -j check".
662    
663    
664    Version 8.34 15-December-2013
665    -----------------------------
666    
667  1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT  1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
668      executable memory. Patch inspired by Carsten Klein.      executable memory. Patch inspired by Carsten Klein.
669    
670  2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although  2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
671      this macro is never tested and has no effect, because the work to support      this macro is never tested and has no effect, because the work to support
672      coverage involves only compiling and linking options and special targets in      coverage involves only compiling and linking options and special targets in
673      the Makefile. The comment in config.h implied that defining the macro would      the Makefile. The comment in config.h implied that defining the macro would
674      enable coverage support, which is totally false. There was also support for      enable coverage support, which is totally false. There was also support for
675      setting this macro in the CMake files (my fault, I just copied it from      setting this macro in the CMake files (my fault, I just copied it from
676      configure). SUPPORT_GCOV has now been removed.      configure). SUPPORT_GCOV has now been removed.
677    
678  3.  Make a small performance improvement in strlen16() and strlen32() in  3.  Make a small performance improvement in strlen16() and strlen32() in
679      pcretest.      pcretest.
680    
681  4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,  4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
682      detected by the Solaris compiler (gcc doesn't seem to be able to diagnose      detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
683      these cases). There was also one in pcretest.c.      these cases). There was also one in pcretest.c.
684    
685  5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.  5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
686    
687  6.  In UTF mode, the code for checking whether a group could match an empty  6.  In UTF mode, the code for checking whether a group could match an empty
688      string (which is used for indefinitely repeated groups to allow for      string (which is used for indefinitely repeated groups to allow for
# Line 31  Version 8.34 xx-xxxx-201x Line 691  Version 8.34 xx-xxxx-201x
691      data item and had a minimum repetition of zero (for example, [^\x{100}]* in      data item and had a minimum repetition of zero (for example, [^\x{100}]* in
692      UTF-8 mode). The effect was undefined: the group might or might not be      UTF-8 mode). The effect was undefined: the group might or might not be
693      deemed as matching an empty string, or the program might have crashed.      deemed as matching an empty string, or the program might have crashed.
694    
695  7.  The code for checking whether a group could match an empty string was not  7.  The code for checking whether a group could match an empty string was not
696      recognizing that \h, \H, \v, \V, and \R must match a character.      recognizing that \h, \H, \v, \V, and \R must match a character.
697    
698  8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match  8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
699      an empty string. If it can, pcretest shows this in its information output.      an empty string. If it can, pcretest shows this in its information output.
700    
701  9.  Fixed two related bugs that applied to Unicode extended grapheme clusters  9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
702      that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when      that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
703      matched by pcre_exec() without using JIT:      matched by pcre_exec() without using JIT:
704    
705      (a) If the rest of the pattern did not match after a maximal run of      (a) If the rest of the pattern did not match after a maximal run of
706          grapheme clusters, the code for backing up to try with fewer of them          grapheme clusters, the code for backing up to try with fewer of them
707          did not always back up over a full grapheme when characters that do not          did not always back up over a full grapheme when characters that do not
708          have the modifier quality were involved, e.g. Hangul syllables.          have the modifier quality were involved, e.g. Hangul syllables.
709    
710      (b) If the match point in a subject started with modifier character, and      (b) If the match point in a subject started with modifier character, and
711          there was no match, the code could incorrectly back up beyond the match          there was no match, the code could incorrectly back up beyond the match
712          point, and potentially beyond the first character in the subject,          point, and potentially beyond the first character in the subject,
713          leading to a segfault or an incorrect match result.          leading to a segfault or an incorrect match result.
714    
715  10. A conditional group with an assertion condition could lead to PCRE  10. A conditional group with an assertion condition could lead to PCRE
716      recording an incorrect first data item for a match if no other first data      recording an incorrect first data item for a match if no other first data
717      item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a      item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
718      first data item, and therefore matched "ca" after "c" instead of at the      first data item, and therefore matched "ca" after "c" instead of at the
719      start.      start.
720    
721  11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a  11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
722      bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.      bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
723    
724  12. The source of pcregrep now includes z/OS-specific code so that it can be  12. The source of pcregrep now includes z/OS-specific code so that it can be
725      compiled for z/OS as part of the special z/OS distribution.      compiled for z/OS as part of the special z/OS distribution.
726    
727  13. Added the -T and -TM options to pcretest.  13. Added the -T and -TM options to pcretest.
728    
729  14. The code in pcre_compile.c for creating the table of named capturing groups  14. The code in pcre_compile.c for creating the table of named capturing groups
730      has been refactored. Instead of creating the table dynamically during the      has been refactored. Instead of creating the table dynamically during the
731      actual compiling pass, the information is remembered during the pre-compile      actual compiling pass, the information is remembered during the pre-compile
732      pass (on the stack unless there are more than 20 named groups, in which      pass (on the stack unless there are more than 20 named groups, in which
733      case malloc() is used) and the whole table is created before the actual      case malloc() is used) and the whole table is created before the actual
734      compile happens. This has simplified the code (it is now nearly 150 lines      compile happens. This has simplified the code (it is now nearly 150 lines
735      shorter) and prepared the way for better handling of references to groups      shorter) and prepared the way for better handling of references to groups
736      with duplicate names.      with duplicate names.
737    
738  15. A back reference to a named subpattern when there is more than one of the  15. A back reference to a named subpattern when there is more than one of the
739      same name now checks them in the order in which they appear in the pattern.      same name now checks them in the order in which they appear in the pattern.
740      The first one that is set is used for the reference. Previously only the      The first one that is set is used for the reference. Previously only the
741      first one was inspected. This change makes PCRE more compatible with Perl.      first one was inspected. This change makes PCRE more compatible with Perl.
742    
743  16. Unicode character properties were updated from Unicode 6.3.0.  16. Unicode character properties were updated from Unicode 6.3.0.
744    
745  17. The compile-time code for auto-possessification has been refactored, based  17. The compile-time code for auto-possessification has been refactored, based
746      on a patch by Zoltan Herczeg. It now happens after instead of during      on a patch by Zoltan Herczeg. It now happens after instead of during
747      compilation. The code is cleaner, and more cases are handled. The option      compilation. The code is cleaner, and more cases are handled. The option
748      PCRE_NO_AUTO_POSSESSIFY is added for testing purposes, and the -O and /O      PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
749      options in pcretest are provided to set it.      options in pcretest are provided to set it. It can also be set by
750        (*NO_AUTO_POSSESS) at the start of a pattern.
751  18. The character VT has been added to the set of characters that match \s and  
752      are generally treated as white space, following this same change in Perl  18. The character VT has been added to the default ("C" locale) set of
753      5.18. There is now no difference between "Perl space" and "POSIX space".      characters that match \s and are generally treated as white space,
754        following this same change in Perl 5.18. There is now no difference between
755  19. The code for checking named groups as conditions, either for being set or      "Perl space" and "POSIX space". Whether VT is treated as white space in
756        other locales depends on the locale.
757    
758    19. The code for checking named groups as conditions, either for being set or
759      for being recursed, has been refactored (this is related to 14 and 15      for being recursed, has been refactored (this is related to 14 and 15
760      above). Processing unduplicated named groups should now be as fast at      above). Processing unduplicated named groups should now be as fast at
761      numerical groups, and processing duplicated groups should be faster than      numerical groups, and processing duplicated groups should be faster than
762      before.      before.
763    
764  20. Two patches to the CMake build system, by Alexander Barkov:  20. Two patches to the CMake build system, by Alexander Barkov:
765    
766        (1) Replace the "source" command by "." in CMakeLists.txt because        (1) Replace the "source" command by "." in CMakeLists.txt because
767            "source" is a bash-ism.            "source" is a bash-ism.
768    
769        (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;        (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
770            without these the CMake build does not work on Solaris.            without these the CMake build does not work on Solaris.
771    
772  21. Perl has changed its handling of \8 and \9. If there is no previously  21. Perl has changed its handling of \8 and \9. If there is no previously
773      encountered capturing group of those numbers, they are treated as the      encountered capturing group of those numbers, they are treated as the
774      literal characters 8 and 9 instead of a binary zero followed by the      literal characters 8 and 9 instead of a binary zero followed by the
775      literals. PCRE now does the same.      literals. PCRE now does the same.
776    
777  22. Following Perl, added \o{} to specify codepoints in octal, making it  22. Following Perl, added \o{} to specify codepoints in octal, making it
778      possible to specify values greater than 0777 and also making them      possible to specify values greater than 0777 and also making them
779      unambiguous.      unambiguous.
780    
781  23. Perl now gives an error for missing closing braces after \x{... instead of  23. Perl now gives an error for missing closing braces after \x{... instead of
782      treating the string as literal. PCRE now does the same.      treating the string as literal. PCRE now does the same.
783    
784  24. RunTest used to grumble if an inappropriate test was selected explicitly,  24. RunTest used to grumble if an inappropriate test was selected explicitly,
785      but just skip it when running all tests. This make it awkward to run ranges      but just skip it when running all tests. This make it awkward to run ranges
786      of tests when one of them was inappropriate. Now it just skips any      of tests when one of them was inappropriate. Now it just skips any
787      inappropriate tests, as it always did when running all tests.      inappropriate tests, as it always did when running all tests.
788    
789  25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained  25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
790      character types such as \d or \w, too many callouts were inserted, and the      character types such as \d or \w, too many callouts were inserted, and the
791      data that they returned was rubbish.      data that they returned was rubbish.
792    
793  26. In UCP mode, \s was not matching two of the characters that Perl matches,  26. In UCP mode, \s was not matching two of the characters that Perl matches,
794      namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they      namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
795      were matched by \h. The code has now been refactored so that the lists of      were matched by \h. The code has now been refactored so that the lists of
796      the horizontal and vertical whitespace characters used for \h and \v (which      the horizontal and vertical whitespace characters used for \h and \v (which
797      are defined only in one place) are now also used for \s.      are defined only in one place) are now also used for \s.
798    
799  27. Add JIT support for the 64 bit TileGX architecture.  27. Add JIT support for the 64 bit TileGX architecture.
800      Patch by Jiong Wang (Tilera Corporation).      Patch by Jiong Wang (Tilera Corporation).
801    
802  28. Possessive quantifiers for classes (both explicit and automatically  28. Possessive quantifiers for classes (both explicit and automatically
803      generated) now use special opcodes instead of wrapping in ONCE brackets.      generated) now use special opcodes instead of wrapping in ONCE brackets.
804    
805  29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier  29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
806      (because it's meaningless), this was not happening when PCRE_CASELESS was      (because it's meaningless), this was not happening when PCRE_CASELESS was
807      set. Not wrong, but inefficient.      set. Not wrong, but inefficient.
808    
809  30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode  30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
810      properties for \w, \d, etc) is present in a test regex. Otherwise if the      properties for \w, \d, etc) is present in a test regex. Otherwise if the
811      test contains no characters greater than 255, Perl doesn't realise it      test contains no characters greater than 255, Perl doesn't realise it
812      should be using Unicode semantics.      should be using Unicode semantics.
813    
814  31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and  31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
815      [:punct:] when PCRE_UCP is set so as to include the same characters as Perl      [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
816      does in Unicode mode.      does in Unicode mode.
817    
818  32. Added the "forbid" facility to pcretest so that putting tests into the  32. Added the "forbid" facility to pcretest so that putting tests into the
819      wrong test files can sometimes be quickly detected.      wrong test files can sometimes be quickly detected.
820    
821    33. There is now a limit (default 250) on the depth of nesting of parentheses.
822        This limit is imposed to control the amount of system stack used at compile
823        time. It can be changed at build time by --with-parens-nest-limit=xxx or
824        the equivalent in CMake.
825    
826    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
827        errors. Perl warns for these when in warning mode, but PCRE has no facility
828        for giving warnings.
829    
830    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
831        However, this was not working for (?!) because it is optimized to (*FAIL),
832        for which PCRE does not allow quantifiers. The optimization is now disabled
833        when a quantifier follows (?!). I can't see any use for this, but it makes
834        things uniform.
835    
836    36. Perl no longer allows group names to start with digits, so I have made this
837        change also in PCRE. It simplifies the code a bit.
838    
839    37. In extended mode, Perl ignores spaces before a + that indicates a
840        possessive quantifier. PCRE allowed a space before the quantifier, but not
841        before the possessive +. It now does.
842    
843    38. The use of \K (reset reported match start) within a repeated possessive
844        group such as (a\Kb)*+ was not working.
845    
846    40. Document that the same character tables must be used at compile time and
847        run time, and that the facility to pass tables to pcre_exec() and
848        pcre_dfa_exec() is for use only with saved/restored patterns.
849    
850    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
851        features for Builds with MSVC:
852    
853        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
854           builds). The .rc files can be used to set FileDescription and many other
855           attributes.
856    
857        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
858           This allows higher-level build scripts which want .pdb files to avoid
859           hard-coding the exact files needed."
860    
861    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
862        mean "start of word" and "end of word", respectively, as a transition aid.
863    
864    43. A minimizing repeat of a class containing codepoints greater than 255 in
865        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
866        compiled to use the heap for recursion.
867    
868    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
869        is configured.
870    
871    
872  Version 8.33 28-May-2013  Version 8.33 28-May-2013
873  ------------------------  ------------------------

Legend:
Removed from v.1388  
changed lines
  Added in v.1658

  ViewVC Help
Powered by ViewVC 1.1.5