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

Diff of /code/trunk/ChangeLog

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

revision 319 by ph10, Mon Jan 28 15:27:45 2008 UTC revision 1691 by ph10, Mon Mar 27 16:00:16 2017 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    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 8.41
8    ------------
9    
10    1.  Fixed typo in CMakeLists.txt (wrong number of arguments for
11    PCRE_STATIC_RUNTIME (affects MSVC only).
12    
13    2.  Issue 1 for 8.40 below was not correctly fixed. If pcregrep in multiline
14    mode with --only-matching matched several lines, it restarted scanning at the
15    next line instead of moving on to the end of the matched string, which can be
16    several lines after the start.
17    
18    3.  Fix a missing else in the JIT compiler reported by 'idaifish'.
19    
20    4.  A (?# style comment is now ignored between a basic quantifier and a
21    following '+' or '?' (example: /X+(?#comment)?Y/.
22    
23    5.  Avoid use of a potentially overflowing buffer in pcregrep (patch by Petr
24    Pisar).
25    
26    6.  Fuzzers have reported issues in pcretest. These are NOT serious (it is,
27    after all, just a test program). However, to stop the reports, some easy ones
28    are fixed:
29    
30        (a) Check for values < 256 when calling isprint() in pcretest.
31        (b) Give an error for too big a number after \O.
32    
33    7.  In the 32-bit library in non-UTF mode, an attempt to find a Unicode
34    property for a character with a code point greater than 0x10ffff (the Unicode
35    maximum) caused a crash.
36    
37    8. The alternative matching function, pcre_dfa_exec() misbehaved if it
38    encountered a character class with a possessive repeat, for example [a-f]{3}+.
39    
40    9. When pcretest called pcre_copy_substring() in 32-bit mode, it set the buffer
41    length incorrectly, which could result in buffer overflow.
42    
43    
44    Version 8.40 11-January-2017
45    ----------------------------
46    
47    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
48        the match extended over a line boundary.
49    
50    2.  Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
51        MSVC static compilation, putting the first patch under a new option.
52    
53    3.  Fix register overwite in JIT when SSE2 acceleration is enabled.
54    
55    4.  Ignore "show all captures" (/=) for DFA matching.
56    
57    5.  Fix JIT unaligned accesses on x86. Patch by Marc Mutz.
58    
59    6.  In any wide-character mode (8-bit UTF or any 16-bit or 32-bit mode),
60        without PCRE_UCP set, a negative character type such as \D in a positive
61        class should cause all characters greater than 255 to match, whatever else
62        is in the class. There was a bug that caused this not to happen if a
63        Unicode property item was added to such a class, for example [\D\P{Nd}] or
64        [\W\pL].
65    
66    7.  When pcretest was outputing information from a callout, the caret indicator
67        for the current position in the subject line was incorrect if it was after
68        an escape sequence for a character whose code point was greater than
69        \x{ff}.
70    
71    8.  A pattern such as (?<RA>abc)(?(R)xyz) was incorrectly compiled such that
72        the conditional was interpreted as a reference to capturing group 1 instead
73        of a test for recursion. Any group whose name began with R was
74        misinterpreted in this way. (The reference interpretation should only
75        happen if the group's name is precisely "R".)
76    
77    9.  A number of bugs have been mended relating to match start-up optimizations
78        when the first thing in a pattern is a positive lookahead. These all
79        applied only when PCRE_NO_START_OPTIMIZE was *not* set:
80    
81        (a) A pattern such as (?=.*X)X$ was incorrectly optimized as if it needed
82            both an initial 'X' and a following 'X'.
83        (b) Some patterns starting with an assertion that started with .* were
84            incorrectly optimized as having to match at the start of the subject or
85            after a newline. There are cases where this is not true, for example,
86            (?=.*[A-Z])(?=.{8,16})(?!.*[\s]) matches after the start in lines that
87            start with spaces. Starting .* in an assertion is no longer taken as an
88            indication of matching at the start (or after a newline).
89    
90    
91    Version 8.39 14-June-2016
92    -------------------------
93    
94    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
95        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
96        misbehaved. This bug was found by the LLVM fuzzer.
97    
98    2.  Similar to the above, if an isolated \E was present between an item and its
99        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
100        bug was found by the LLVM fuzzer.
101    
102    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
103        working correctly in UCP mode.
104    
105    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
106        was set when the pmatch argument was NULL. It now returns REG_INVARG.
107    
108    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
109    
110    6.  An empty \Q\E sequence between an item and its qualifier caused
111        pcre_compile() to misbehave when auto callouts were enabled. This bug was
112        found by the LLVM fuzzer.
113    
114    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
115        space or a #-type comment that was followed by (?-x), which turns off
116        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
117        pcre_compile() assumed that (?-x) applied to the whole pattern and
118        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
119    
120    8.  A call of pcre_copy_named_substring() for a named substring whose number
121        was greater than the space in the ovector could cause a crash.
122    
123    9.  Yet another buffer overflow bug involved duplicate named groups with a
124        group that reset capture numbers (compare 8.38/7 below). Once again, I have
125        just allowed for more memory, even if not needed. (A proper fix is
126        implemented in PCRE2, but it involves a lot of refactoring.)
127    
128    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
129        start of the match to be earlier than the end.
130    
131    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
132    
133    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
134        assertion, caused pcretest to generate incorrect output, and also to read
135        uninitialized memory (detected by ASAN or valgrind).
136    
137    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
138        nested set of parentheses of sufficient size caused an overflow of the
139        compiling workspace (which was diagnosed, but of course is not desirable).
140    
141    14. And yet another buffer overflow bug involving duplicate named groups, this
142        time nested, with a nested back reference. Yet again, I have just allowed
143        for more memory, because anything more needs all the refactoring that has
144        been done for PCRE2. An example pattern that provoked this bug is:
145        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
146        registered as CVE-2016-1283.
147    
148    15. pcretest went into a loop if global matching was requested with an ovector
149        size less than 2. It now gives an error message. This bug was found by
150        afl-fuzz.
151    
152    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
153        ("assertion expected") when (?(?C) was not followed by an opening
154        parenthesis.
155    
156    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
157        actually affect anything, by sheer luck.
158    
159    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
160        static compilation.
161    
162    19. Modified the RunTest script to incorporate a valgrind suppressions file so
163        that certain errors, provoked by the SSE2 instruction set when JIT is used,
164        are ignored.
165    
166    20. A racing condition is fixed in JIT reported by Mozilla.
167    
168    21. Minor code refactor to avoid "array subscript is below array bounds"
169        compiler warning.
170    
171    22. Minor code refactor to avoid "left shift of negative number" warning.
172    
173    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
174        UCP support.
175    
176    24. Refactor to avoid compiler warnings in pcrecpp.cc.
177    
178    25. Refactor to fix a typo in pcre_jit_test.c
179    
180    26. Patch to support compiling pcrecpp.cc with Intel compiler.
181    
182    
183    Version 8.38 23-November-2015
184    -----------------------------
185    
186    1.  If a group that contained a recursive back reference also contained a
187        forward reference subroutine call followed by a non-forward-reference
188        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
189        compile correct code, leading to undefined behaviour or an internally
190        detected error. This bug was discovered by the LLVM fuzzer.
191    
192    2.  Quantification of certain items (e.g. atomic back references) could cause
193        incorrect code to be compiled when recursive forward references were
194        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
195        This bug was discovered by the LLVM fuzzer.
196    
197    3.  A repeated conditional group whose condition was a reference by name caused
198        a buffer overflow if there was more than one group with the given name.
199        This bug was discovered by the LLVM fuzzer.
200    
201    4.  A recursive back reference by name within a group that had the same name as
202        another group caused a buffer overflow. For example:
203        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
204    
205    5.  A forward reference by name to a group whose number is the same as the
206        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
207        a buffer overflow at compile time. This bug was discovered by the LLVM
208        fuzzer.
209    
210    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
211        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
212    
213    7.  Another buffer overflow bug involved duplicate named groups with a
214        reference between their definition, with a group that reset capture
215        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
216        fixed by always allowing for more memory, even if not needed. (A proper fix
217        is implemented in PCRE2, but it involves more refactoring.)
218    
219    8.  There was no check for integer overflow in subroutine calls such as (?123).
220    
221    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
222        being treated as a literal 'l' instead of causing an error.
223    
224    10. There was a buffer overflow if pcre_exec() was called with an ovector of
225        size 1. This bug was found by american fuzzy lop.
226    
227    11. If a non-capturing group containing a conditional group that could match
228        an empty string was repeated, it was not identified as matching an empty
229        string itself. For example: /^(?:(?(1)x|)+)+$()/.
230    
231    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
232        \a and \e in test subject lines.
233    
234    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
235        instead of the EBCDIC value.
236    
237    14. The handling of \c in an EBCDIC environment has been revised so that it is
238        now compatible with the specification in Perl's perlebcdic page.
239    
240    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
241        ASCII/Unicode. This has now been added to the list of characters that are
242        recognized as white space in EBCDIC.
243    
244    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
245        error (correctly) when used outside a class, but did not give an error
246        within a class.
247    
248    17. \h within a class was incorrectly compiled in EBCDIC environments.
249    
250    18. A pattern with an unmatched closing parenthesis that contained a backward
251        assertion which itself contained a forward reference caused buffer
252        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
253    
254    19. JIT should return with error when the compiled pattern requires more stack
255        space than the maximum.
256    
257    20. A possessively repeated conditional group that could match an empty string,
258        for example, /(?(R))*+/, was incorrectly compiled.
259    
260    21. Fix infinite recursion in the JIT compiler when certain patterns such as
261        /(?:|a|){100}x/ are analysed.
262    
263    22. Some patterns with character classes involving [: and \\ were incorrectly
264        compiled and could cause reading from uninitialized memory or an incorrect
265        error diagnosis.
266    
267    23. Pathological patterns containing many nested occurrences of [: caused
268        pcre_compile() to run for a very long time.
269    
270    24. A conditional group with only one branch has an implicit empty alternative
271        branch and must therefore be treated as potentially matching an empty
272        string.
273    
274    25. If (?R was followed by - or + incorrect behaviour happened instead of a
275        diagnostic.
276    
277    26. Arrange to give up on finding the minimum matching length for overly
278        complex patterns.
279    
280    27. Similar to (4) above: in a pattern with duplicated named groups and an
281        occurrence of (?| it is possible for an apparently non-recursive back
282        reference to become recursive if a later named group with the relevant
283        number is encountered. This could lead to a buffer overflow. Wen Guanxing
284        from Venustech ADLAB discovered this bug.
285    
286    28. If pcregrep was given the -q option with -c or -l, or when handling a
287        binary file, it incorrectly wrote output to stdout.
288    
289    29. The JIT compiler did not restore the control verb head in case of *THEN
290        control verbs. This issue was found by Karl Skomski with a custom LLVM
291        fuzzer.
292    
293    30. Error messages for syntax errors following \g and \k were giving inaccurate
294        offsets in the pattern.
295    
296    31. Added a check for integer overflow in conditions (?(<digits>) and
297        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
298        fuzzer.
299    
300    32. Handling recursive references such as (?2) when the reference is to a group
301        later in the pattern uses code that is very hacked about and error-prone.
302        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
303        give an internal error if it is obvious that compiling has gone wrong.
304    
305    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
306        This issue was found by Karl Skomski with a custom LLVM fuzzer.
307    
308    34. The JIT compiler should restore the control chain for empty possessive
309        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
310    
311    35. Match limit check added to JIT recursion. This issue was found by Karl
312        Skomski with a custom LLVM fuzzer.
313    
314    36. Yet another case similar to 27 above has been circumvented by an
315        unconditional allocation of extra memory. This issue is fixed "properly" in
316        PCRE2 by refactoring the way references are handled. Wen Guanxing
317        from Venustech ADLAB discovered this bug.
318    
319    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
320        with a custom LLVM fuzzer.
321    
322    38. Fixed a corner case of range optimization in JIT.
323    
324    39. An incorrect error "overran compiling workspace" was given if there were
325        exactly enough group forward references such that the last one extended
326        into the workspace safety margin. The next one would have expanded the
327        workspace. The test for overflow was not including the safety margin.
328    
329    40. A match limit issue is fixed in JIT which was found by Karl Skomski
330        with a custom LLVM fuzzer.
331    
332    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
333        work under Windows. (Why has it taken so long for anyone to notice?)
334    
335    42. In a character class such as [\W\p{Any}] where both a negative-type escape
336        ("not a word character") and a property escape were present, the property
337        escape was being ignored.
338    
339    43. Fix crash caused by very long (*MARK) or (*THEN) names.
340    
341    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
342        by a single ASCII character in a class item, was incorrectly compiled in
343        UCP mode. The POSIX class got lost, but only if the single character
344        followed it.
345    
346    45. [:punct:] in UCP mode was matching some characters in the range 128-255
347        that should not have been matched.
348    
349    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
350        class, all characters with code points greater than 255 are in the class.
351        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
352        such as \w are turned into Unicode properties), wide characters were not
353        correctly handled, and could fail to match.
354    
355    
356    Version 8.37 28-April-2015
357    --------------------------
358    
359    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
360        for those parentheses to be closed with whatever has been captured so far.
361        However, it was failing to mark any other groups between the hightest
362        capture so far and the currrent group as "unset". Thus, the ovector for
363        those groups contained whatever was previously there. An example is the
364        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
365    
366    2.  If an assertion condition was quantified with a minimum of zero (an odd
367        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
368    
369    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
370        unrecognized modifier, a crash could occur.
371    
372    4.  An attempt to do global matching in pcretest with a zero-length ovector
373        caused a crash.
374    
375    5.  Fixed a memory leak during matching that could occur for a subpattern
376        subroutine call (recursive or otherwise) if the number of captured groups
377        that had to be saved was greater than ten.
378    
379    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
380        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
381        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
382    
383    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
384        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
385    
386    8. When a pattern is compiled, it remembers the highest back reference so that
387       when matching, if the ovector is too small, extra memory can be obtained to
388       use instead. A conditional subpattern whose condition is a check on a
389       capture having happened, such as, for example in the pattern
390       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
391       setting the highest backreference number. This mattered only if pcre_exec()
392       was called with an ovector that was too small to hold the capture, and there
393       was no other kind of back reference (a situation which is probably quite
394       rare). The effect of the bug was that the condition was always treated as
395       FALSE when the capture could not be consulted, leading to a incorrect
396       behaviour by pcre_exec(). This bug has been fixed.
397    
398    9. A reference to a duplicated named group (either a back reference or a test
399       for being set in a conditional) that occurred in a part of the pattern where
400       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
401       to be incorrectly calculated, leading to overwriting.
402    
403    10. A mutually recursive set of back references such as (\2)(\1) caused a
404        segfault at study time (while trying to find the minimum matching length).
405        The infinite loop is now broken (with the minimum length unset, that is,
406        zero).
407    
408    11. If an assertion that was used as a condition was quantified with a minimum
409        of zero, matching went wrong. In particular, if the whole group had
410        unlimited repetition and could match an empty string, a segfault was
411        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
412        assertions to be quantified, but not if they are being used as conditions,
413        so the above pattern is faulted by Perl. PCRE has now been changed so that
414        it also rejects such patterns.
415    
416    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
417        failed to allow the zero-repeat case if pcre2_exec() was called with an
418        ovector too small to capture the group.
419    
420    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
421        Red Hat Product Security:
422    
423        (a) A crash if /K and /F were both set with the option to save the compiled
424        pattern.
425    
426        (b) Another crash if the option to print captured substrings in a callout
427        was combined with setting a null ovector, for example \O\C+ as a subject
428        string.
429    
430    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
431        forward reference repeated a large (but limited) number of times within a
432        repeated outer group that has a zero minimum quantifier, caused incorrect
433        code to be compiled, leading to the error "internal error:
434        previously-checked referenced subpattern not found" when an incorrect
435        memory address was read. This bug was reported as "heap overflow",
436        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
437        CVE-2015-2325.
438    
439    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
440        call within a group that also contained a recursive back reference caused
441        incorrect code to be compiled. This bug was reported as "heap overflow",
442        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
443        number CVE-2015-2326.
444    
445    24. Computing the size of the JIT read-only data in advance has been a source
446        of various issues, and new ones are still appear unfortunately. To fix
447        existing and future issues, size computation is eliminated from the code,
448        and replaced by on-demand memory allocation.
449    
450    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
451        adjacent to the end of the range, and the range contained characters with
452        more than one other case, caused incorrect behaviour when compiled in UTF
453        mode. In that example, the range a-j was left out of the class.
454    
455    26. Fix JIT compilation of conditional blocks, which assertion
456        is converted to (*FAIL). E.g: /(?(?!))/.
457    
458    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
459        discovered by the LLVM fuzzer.
460    
461    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
462        when this assertion was used as a condition, for example (?(?!)a|b). In
463        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
464        error about an unsupported item.
465    
466    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
467        possessification code could take exponential time to complete. A recursion
468        depth limit of 1000 has been imposed to limit the resources used by this
469        optimization.
470    
471    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
472        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
473        because \S ensures they are all in the class. The code for doing this was
474        interacting badly with the code for computing the amount of space needed to
475        compile the pattern, leading to a buffer overflow. This bug was discovered
476        by the LLVM fuzzer.
477    
478    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
479        other kinds of group caused stack overflow at compile time. This bug was
480        discovered by the LLVM fuzzer.
481    
482    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
483        between a subroutine call and its quantifier was incorrectly compiled,
484        leading to buffer overflow or other errors. This bug was discovered by the
485        LLVM fuzzer.
486    
487    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
488        assertion after (?(. The code was failing to check the character after
489        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
490        was discovered by the LLVM fuzzer.
491    
492    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
493        a fixed maximum following a group that contains a subroutine reference was
494        incorrectly compiled and could trigger buffer overflow. This bug was
495        discovered by the LLVM fuzzer.
496    
497    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
498        caused a stack overflow instead of the diagnosis of a non-fixed length
499        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
500    
501    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
502        (e.g. /(?<=\Ka)/) could make pcregrep loop.
503    
504    37. There was a similar problem to 36 in pcretest for global matches.
505    
506    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
507        and a subsequent item in the pattern caused a non-match, backtracking over
508        the repeated \X did not stop, but carried on past the start of the subject,
509        causing reference to random memory and/or a segfault. There were also some
510        other cases where backtracking after \C could crash. This set of bugs was
511        discovered by the LLVM fuzzer.
512    
513    39. The function for finding the minimum length of a matching string could take
514        a very long time if mutual recursion was present many times in a pattern,
515        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
516        method has been implemented. This infelicity was discovered by the LLVM
517        fuzzer.
518    
519    40. Static linking against the PCRE library using the pkg-config module was
520        failing on missing pthread symbols.
521    
522    
523    Version 8.36 26-September-2014
524    ------------------------------
525    
526    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
527        -Wmissing-field-initializers and -Wunused-parameter.
528    
529    2.  The tests for quantifiers being too big (greater than 65535) were being
530        applied after reading the number, and stupidly assuming that integer
531        overflow would give a negative number. The tests are now applied as the
532        numbers are read.
533    
534    3.  Tidy code in pcre_exec.c where two branches that used to be different are
535        now the same.
536    
537    4.  The JIT compiler did not generate match limit checks for certain
538        bracketed expressions with quantifiers. This may lead to exponential
539        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
540        issue should be resolved now.
541    
542    5.  Fixed an issue, which occures when nested alternatives are optimized
543        with table jumps.
544    
545    6.  Inserted two casts and changed some ints to size_t in the light of some
546        reported 64-bit compiler warnings (Bugzilla 1477).
547    
548    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
549        an empty string, which sometimes were behaving incorrectly in the
550        interpreter (though correctly in the JIT matcher). This pcretest input is
551        an example:
552    
553          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
554          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
555    
556        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
557        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
558        for an empty string was breaking the inner loop and carrying on at a lower
559        level, when possessive repeated groups should always return to a higher
560        level as they have no backtrack points in them. The empty string test now
561        occurs at the outer level.
562    
563    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
564        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
565    
566    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
567        doesn't).
568    
569    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
570        to hold a pattern to be incorrectly computed (too small) when there were
571        named back references to duplicated names. This could cause "internal
572        error: code overflow" or "double free or corruption" or other memory
573        handling errors.
574    
575    11. When named subpatterns had the same prefixes, back references could be
576        confused. For example, in this pattern:
577    
578          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
579    
580        the reference to 'Name' was incorrectly treated as a reference to a
581        duplicate name.
582    
583    12. A pattern such as /^s?c/mi8 where the optional character has more than
584        one "other case" was incorrectly compiled such that it would only try to
585        match starting at "c".
586    
587    13. When a pattern starting with \s was studied, VT was not included in the
588        list of possible starting characters; this should have been part of the
589        8.34/18 patch.
590    
591    14. If a character class started [\Qx]... where x is any character, the class
592        was incorrectly terminated at the ].
593    
594    15. If a pattern that started with a caseless match for a character with more
595        than one "other case" was studied, PCRE did not set up the starting code
596        unit bit map for the list of possible characters. Now it does. This is an
597        optimization improvement, not a bug fix.
598    
599    16. The Unicode data tables have been updated to Unicode 7.0.0.
600    
601    17. Fixed a number of memory leaks in pcregrep.
602    
603    18. Avoid a compiler warning (from some compilers) for a function call with
604        a cast that removes "const" from an lvalue by using an intermediate
605        variable (to which the compiler does not object).
606    
607    19. Incorrect code was compiled if a group that contained an internal recursive
608        back reference was optional (had quantifier with a minimum of zero). This
609        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
610        caused segmentation faults because of stack overflows at compile time.
611    
612    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
613        group that is quantified with an indefinite repeat, caused a compile-time
614        loop which used up all the system stack and provoked a segmentation fault.
615        This was not the same bug as 19 above.
616    
617    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
618        Patch by Mike Frysinger.
619    
620    
621    Version 8.35 04-April-2014
622    --------------------------
623    
624    1.  A new flag is set, when property checks are present in an XCLASS.
625        When this flag is not set, PCRE can perform certain optimizations
626        such as studying these XCLASS-es.
627    
628    2.  The auto-possessification of character sets were improved: a normal
629        and an extended character set can be compared now. Furthermore
630        the JIT compiler optimizes more character set checks.
631    
632    3.  Got rid of some compiler warnings for potentially uninitialized variables
633        that show up only when compiled with -O2.
634    
635    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
636        of a match later then the end of the match. The pcretest program was not
637        handling the case sensibly - it was outputting from the start to the next
638        binary zero. It now reports this situation in a message, and outputs the
639        text from the end to the start.
640    
641    5.  Fast forward search is improved in JIT. Instead of the first three
642        characters, any three characters with fixed position can be searched.
643        Search order: first, last, middle.
644    
645    6.  Improve character range checks in JIT. Characters are read by an inprecise
646        function now, which returns with an unknown value if the character code is
647        above a certain threshold (e.g: 256). The only limitation is that the value
648        must be bigger than the threshold as well. This function is useful when
649        the characters above the threshold are handled in the same way.
650    
651    7.  The macros whose names start with RAWUCHAR are placeholders for a future
652        mode in which only the bottom 21 bits of 32-bit data items are used. To
653        make this more memorable for those maintaining the code, the names have
654        been changed to start with UCHAR21, and an extensive comment has been added
655        to their definition.
656    
657    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
658        to the export list in Makefile.am (they were accidentally omitted from the
659        8.34 tarball).
660    
661    9.  The informational output from pcretest used the phrase "starting byte set"
662        which is inappropriate for the 16-bit and 32-bit libraries. As the output
663        for "first char" and "need char" really means "non-UTF-char", I've changed
664        "byte" to "char", and slightly reworded the output. The documentation about
665        these values has also been (I hope) clarified.
666    
667    10. Another JIT related optimization: use table jumps for selecting the correct
668        backtracking path, when more than four alternatives are present inside a
669        bracket.
670    
671    11. Empty match is not possible, when the minimum length is greater than zero,
672        and there is no \K in the pattern. JIT should avoid empty match checks in
673        such cases.
674    
675    12. In a caseless character class with UCP support, when a character with more
676        than one alternative case was not the first character of a range, not all
677        the alternative cases were added to the class. For example, s and \x{17f}
678        are both alternative cases for S: the class [RST] was handled correctly,
679        but [R-T] was not.
680    
681    13. The configure.ac file always checked for pthread support when JIT was
682        enabled. This is not used in Windows, so I have put this test inside a
683        check for the presence of windows.h (which was already tested for).
684    
685    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
686        The algorithm provides a way to skip certain starting offsets, and usually
687        faster than linear prefix searches.
688    
689    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
690        as for 'fr_FR' and 'french'. For some reason, however, it then used the
691        Windows-specific input and output files, which have 'french' screwed in.
692        So this could never have worked. One of the problems with locales is that
693        they aren't always the same. I have now updated RunTest so that it checks
694        the output of the locale test (test 3) against three different output
695        files, and it allows the test to pass if any one of them matches. With luck
696        this should make the test pass on some versions of Solaris where it was
697        failing. Because of the uncertainty, the script did not used to stop if
698        test 3 failed; it now does. If further versions of a French locale ever
699        come to light, they can now easily be added.
700    
701    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
702        there was a message during ./configure, but it did not stop. This now
703        provokes an error. The invalid example in README has been corrected.
704        If a value less than the minimum is given, the minimum value has always
705        been used, but now a warning is given.
706    
707    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
708        was a bug in the test system, which is now fixed. Also, the list of various
709        configurations that are tested for each release did not have one with both
710        16/32 bits and --enable-bar-anycrlf. It now does.
711    
712    18. pcretest was missing "-C bsr" for displaying the \R default setting.
713    
714    19. Little endian PowerPC systems are supported now by the JIT compiler.
715    
716    20. The fast forward newline mechanism could enter to an infinite loop on
717        certain invalid UTF-8 input. Although we don't support these cases
718        this issue can be fixed by a performance optimization.
719    
720    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
721        not take account if existing stack usage. There is now a new global
722        variable called pcre_stack_guard that can be set to point to an external
723        function to check stack availability. It is called at the start of
724        processing every parenthesized group.
725    
726    22. A typo in the code meant that in ungreedy mode the max/min qualifier
727        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
728        not match "ab".
729    
730    23. When UTF was disabled, the JIT program reported some incorrect compile
731        errors. These messages are silenced now.
732    
733    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
734        compiler.
735    
736    25. Change all the temporary files used in RunGrepTest to be different to those
737        used by RunTest so that the tests can be run simultaneously, for example by
738        "make -j check".
739    
740    
741    Version 8.34 15-December-2013
742    -----------------------------
743    
744    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
745        executable memory. Patch inspired by Carsten Klein.
746    
747    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
748        this macro is never tested and has no effect, because the work to support
749        coverage involves only compiling and linking options and special targets in
750        the Makefile. The comment in config.h implied that defining the macro would
751        enable coverage support, which is totally false. There was also support for
752        setting this macro in the CMake files (my fault, I just copied it from
753        configure). SUPPORT_GCOV has now been removed.
754    
755    3.  Make a small performance improvement in strlen16() and strlen32() in
756        pcretest.
757    
758    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
759        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
760        these cases). There was also one in pcretest.c.
761    
762    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
763    
764    6.  In UTF mode, the code for checking whether a group could match an empty
765        string (which is used for indefinitely repeated groups to allow for
766        breaking an infinite loop) was broken when the group contained a repeated
767        negated single-character class with a character that occupied more than one
768        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
769        UTF-8 mode). The effect was undefined: the group might or might not be
770        deemed as matching an empty string, or the program might have crashed.
771    
772    7.  The code for checking whether a group could match an empty string was not
773        recognizing that \h, \H, \v, \V, and \R must match a character.
774    
775    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
776        an empty string. If it can, pcretest shows this in its information output.
777    
778    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
779        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
780        matched by pcre_exec() without using JIT:
781    
782        (a) If the rest of the pattern did not match after a maximal run of
783            grapheme clusters, the code for backing up to try with fewer of them
784            did not always back up over a full grapheme when characters that do not
785            have the modifier quality were involved, e.g. Hangul syllables.
786    
787        (b) If the match point in a subject started with modifier character, and
788            there was no match, the code could incorrectly back up beyond the match
789            point, and potentially beyond the first character in the subject,
790            leading to a segfault or an incorrect match result.
791    
792    10. A conditional group with an assertion condition could lead to PCRE
793        recording an incorrect first data item for a match if no other first data
794        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
795        first data item, and therefore matched "ca" after "c" instead of at the
796        start.
797    
798    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
799        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
800    
801    12. The source of pcregrep now includes z/OS-specific code so that it can be
802        compiled for z/OS as part of the special z/OS distribution.
803    
804    13. Added the -T and -TM options to pcretest.
805    
806    14. The code in pcre_compile.c for creating the table of named capturing groups
807        has been refactored. Instead of creating the table dynamically during the
808        actual compiling pass, the information is remembered during the pre-compile
809        pass (on the stack unless there are more than 20 named groups, in which
810        case malloc() is used) and the whole table is created before the actual
811        compile happens. This has simplified the code (it is now nearly 150 lines
812        shorter) and prepared the way for better handling of references to groups
813        with duplicate names.
814    
815    15. A back reference to a named subpattern when there is more than one of the
816        same name now checks them in the order in which they appear in the pattern.
817        The first one that is set is used for the reference. Previously only the
818        first one was inspected. This change makes PCRE more compatible with Perl.
819    
820    16. Unicode character properties were updated from Unicode 6.3.0.
821    
822    17. The compile-time code for auto-possessification has been refactored, based
823        on a patch by Zoltan Herczeg. It now happens after instead of during
824        compilation. The code is cleaner, and more cases are handled. The option
825        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
826        options in pcretest are provided to set it. It can also be set by
827        (*NO_AUTO_POSSESS) at the start of a pattern.
828    
829    18. The character VT has been added to the default ("C" locale) set of
830        characters that match \s and are generally treated as white space,
831        following this same change in Perl 5.18. There is now no difference between
832        "Perl space" and "POSIX space". Whether VT is treated as white space in
833        other locales depends on the locale.
834    
835    19. The code for checking named groups as conditions, either for being set or
836        for being recursed, has been refactored (this is related to 14 and 15
837        above). Processing unduplicated named groups should now be as fast at
838        numerical groups, and processing duplicated groups should be faster than
839        before.
840    
841    20. Two patches to the CMake build system, by Alexander Barkov:
842    
843          (1) Replace the "source" command by "." in CMakeLists.txt because
844              "source" is a bash-ism.
845    
846          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
847              without these the CMake build does not work on Solaris.
848    
849    21. Perl has changed its handling of \8 and \9. If there is no previously
850        encountered capturing group of those numbers, they are treated as the
851        literal characters 8 and 9 instead of a binary zero followed by the
852        literals. PCRE now does the same.
853    
854    22. Following Perl, added \o{} to specify codepoints in octal, making it
855        possible to specify values greater than 0777 and also making them
856        unambiguous.
857    
858    23. Perl now gives an error for missing closing braces after \x{... instead of
859        treating the string as literal. PCRE now does the same.
860    
861    24. RunTest used to grumble if an inappropriate test was selected explicitly,
862        but just skip it when running all tests. This make it awkward to run ranges
863        of tests when one of them was inappropriate. Now it just skips any
864        inappropriate tests, as it always did when running all tests.
865    
866    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
867        character types such as \d or \w, too many callouts were inserted, and the
868        data that they returned was rubbish.
869    
870    26. In UCP mode, \s was not matching two of the characters that Perl matches,
871        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
872        were matched by \h. The code has now been refactored so that the lists of
873        the horizontal and vertical whitespace characters used for \h and \v (which
874        are defined only in one place) are now also used for \s.
875    
876    27. Add JIT support for the 64 bit TileGX architecture.
877        Patch by Jiong Wang (Tilera Corporation).
878    
879    28. Possessive quantifiers for classes (both explicit and automatically
880        generated) now use special opcodes instead of wrapping in ONCE brackets.
881    
882    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
883        (because it's meaningless), this was not happening when PCRE_CASELESS was
884        set. Not wrong, but inefficient.
885    
886    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
887        properties for \w, \d, etc) is present in a test regex. Otherwise if the
888        test contains no characters greater than 255, Perl doesn't realise it
889        should be using Unicode semantics.
890    
891    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
892        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
893        does in Unicode mode.
894    
895    32. Added the "forbid" facility to pcretest so that putting tests into the
896        wrong test files can sometimes be quickly detected.
897    
898    33. There is now a limit (default 250) on the depth of nesting of parentheses.
899        This limit is imposed to control the amount of system stack used at compile
900        time. It can be changed at build time by --with-parens-nest-limit=xxx or
901        the equivalent in CMake.
902    
903    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
904        errors. Perl warns for these when in warning mode, but PCRE has no facility
905        for giving warnings.
906    
907    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
908        However, this was not working for (?!) because it is optimized to (*FAIL),
909        for which PCRE does not allow quantifiers. The optimization is now disabled
910        when a quantifier follows (?!). I can't see any use for this, but it makes
911        things uniform.
912    
913    36. Perl no longer allows group names to start with digits, so I have made this
914        change also in PCRE. It simplifies the code a bit.
915    
916    37. In extended mode, Perl ignores spaces before a + that indicates a
917        possessive quantifier. PCRE allowed a space before the quantifier, but not
918        before the possessive +. It now does.
919    
920    38. The use of \K (reset reported match start) within a repeated possessive
921        group such as (a\Kb)*+ was not working.
922    
923    40. Document that the same character tables must be used at compile time and
924        run time, and that the facility to pass tables to pcre_exec() and
925        pcre_dfa_exec() is for use only with saved/restored patterns.
926    
927    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
928        features for Builds with MSVC:
929    
930        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
931           builds). The .rc files can be used to set FileDescription and many other
932           attributes.
933    
934        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
935           This allows higher-level build scripts which want .pdb files to avoid
936           hard-coding the exact files needed."
937    
938    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
939        mean "start of word" and "end of word", respectively, as a transition aid.
940    
941    43. A minimizing repeat of a class containing codepoints greater than 255 in
942        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
943        compiled to use the heap for recursion.
944    
945    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
946        is configured.
947    
948    
949    Version 8.33 28-May-2013
950    ------------------------
951    
952    1.  Added 'U' to some constants that are compared to unsigned integers, to
953        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
954        variables that are added to signed variables, to ensure the result is
955        signed and can be negated.
956    
957    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
958        CMake config files.
959    
960    3.  Revise the creation of config.h.generic so that all boolean macros are
961        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
962        overriding via -D on the command line possible.
963    
964    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
965        to unsigned int is reported to make a quite noticeable speed difference in
966        a specific Windows environment. Testing on Linux did also appear to show
967        some benefit (and it is clearly not harmful). Also fixed the definition of
968        Xop which should be unsigned.
969    
970    5.  Related to (4), changing the definition of the intermediate variable cc
971        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
972        improvements.
973    
974    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
975        unnecessary spaces.
976    
977    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
978        and later.
979    
980    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
981    
982    9.  Optimizing fast_forward_start_bits in JIT.
983    
984    10. Adding support for callouts in JIT, and fixing some issues revealed
985        during this work. Namely:
986    
987        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
988    
989        (b) Minimum length was not checked before the matching is started.
990    
991    11. The value of capture_last that is passed to callouts was incorrect in some
992        cases when there was a capture on one path that was subsequently abandoned
993        after a backtrack. Also, the capture_last value is now reset after a
994        recursion, since all captures are also reset in this case.
995    
996    12. The interpreter no longer returns the "too many substrings" error in the
997        case when an overflowing capture is in a branch that is subsequently
998        abandoned after a backtrack.
999    
1000    13. In the pathological case when an offset vector of size 2 is used, pcretest
1001        now prints out the matched string after a yield of 0 or 1.
1002    
1003    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
1004        Only supported by the JIT compiler at the moment.
1005    
1006    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
1007    
1008    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
1009        offset of the starting point of the matching process, provided the offsets
1010        vector is large enough.
1011    
1012    17. The \A escape now records a lookbehind value of 1, though its execution
1013        does not actually inspect the previous character. This is to ensure that,
1014        in partial multi-segment matching, at least one character from the old
1015        segment is retained when a new segment is processed. Otherwise, if there
1016        are no lookbehinds in the pattern, \A might match incorrectly at the start
1017        of a new segment.
1018    
1019    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
1020    
1021    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
1022        gives some modest performance improvement in 8-bit mode.
1023    
1024    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
1025        be expressed in certain programming languages using Universal Character
1026        Names.
1027    
1028    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
1029        which points out that "non characters" are not "characters that may not
1030        appear in Unicode strings" but rather "characters that are reserved for
1031        internal use and have only local meaning".
1032    
1033    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
1034        there was a conditional group that depended on an assertion, if the
1035        assertion was false, the callout that immediately followed the alternation
1036        in the condition was skipped when pcre_exec() was used for matching.
1037    
1038    23. Allow an explicit callout to be inserted before an assertion that is the
1039        condition for a conditional group, for compatibility with automatic
1040        callouts, which always insert a callout at this point.
1041    
1042    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
1043        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
1044    
1045    25. (*PRUNE) is now supported by the JIT compiler.
1046    
1047    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
1048    
1049    27. Fix the case where there are two or more SKIPs with arguments that may be
1050        ignored.
1051    
1052    28. (*SKIP) is now supported by the JIT compiler.
1053    
1054    29. (*THEN) is now supported by the JIT compiler.
1055    
1056    30. Update RunTest with additional test selector options.
1057    
1058    31. The way PCRE handles backtracking verbs has been changed in two ways.
1059    
1060        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
1061        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
1062        backtracking. In some cases this makes it more Perl-compatible, but Perl's
1063        rather obscure rules do not always do the same thing.
1064    
1065        (2) Previously, backtracking verbs were confined within assertions. This is
1066        no longer the case for positive assertions, except for (*ACCEPT). Again,
1067        this sometimes improves Perl compatibility, and sometimes does not.
1068    
1069    32. A number of tests that were in test 2 because Perl did things differently
1070        have been moved to test 1, because either Perl or PCRE has changed, and
1071        these tests are now compatible.
1072    
1073    32. Backtracking control verbs are now handled in the same way in JIT and
1074        interpreter.
1075    
1076    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1077        contained a forward subroutine reference caused a compile error.
1078    
1079    34. Auto-detect and optimize limited repetitions in JIT.
1080    
1081    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1082        blocking (*UTF) etc.
1083    
1084    36. In the interpreter, maximizing pattern repetitions for characters and
1085        character types now use tail recursion, which reduces stack usage.
1086    
1087    37. The value of the max lookbehind was not correctly preserved if a compiled
1088        and saved regex was reloaded on a host of different endianness.
1089    
1090    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1091        of the compiled pattern block, expand the flags field from 16 to 32 bits
1092        because it was almost full.
1093    
1094    39. Try madvise first before posix_madvise.
1095    
1096    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1097        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1098        original change locked out all matches of empty strings. This has been
1099        changed so that one match of an empty string per line is recognized.
1100        Subsequent searches on the same line (for colouring or for --only-matching,
1101        for example) do not recognize empty strings.
1102    
1103    41. Applied a user patch to fix a number of spelling mistakes in comments.
1104    
1105    42. Data lines longer than 65536 caused pcretest to crash.
1106    
1107    43. Clarified the data type for length and startoffset arguments for pcre_exec
1108        and pcre_dfa_exec in the function-specific man pages, where they were
1109        explicitly stated to be in bytes, never having been updated. I also added
1110        some clarification to the pcreapi man page.
1111    
1112    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1113        a segmentation fault.
1114    
1115    
1116    Version 8.32 30-November-2012
1117    -----------------------------
1118    
1119    1.  Improved JIT compiler optimizations for first character search and single
1120        character iterators.
1121    
1122    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1123        Patch by Daniel Richard G.
1124    
1125    3.  Single character iterator optimizations in the JIT compiler.
1126    
1127    4.  Improved JIT compiler optimizations for character ranges.
1128    
1129    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1130        Reported by Giuseppe D'Angelo.
1131    
1132    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1133        of a line, was being set incorrectly in cases where .* appeared inside
1134        atomic brackets at the start of a pattern, or where there was a subsequent
1135        *PRUNE or *SKIP.
1136    
1137    7.  Improved instruction cache flush for POWER/PowerPC.
1138        Patch by Daniel Richard G.
1139    
1140    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1141        grep:
1142    
1143        (a) There is now no limit to the number of patterns to be matched.
1144    
1145        (b) An error is given if a pattern is too long.
1146    
1147        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1148            are now supported.
1149    
1150        (d) --exclude-from and --include-from (multiple use) have been added.
1151    
1152        (e) Exclusions and inclusions now apply to all files and directories, not
1153            just to those obtained from scanning a directory recursively.
1154    
1155        (f) Multiple uses of -f and --file-list are now supported.
1156    
1157        (g) In a Windows environment, the default for -d has been changed from
1158            "read" (the GNU grep default) to "skip", because otherwise the presence
1159            of a directory in the file list provokes an error.
1160    
1161        (h) The documentation has been revised and clarified in places.
1162    
1163    9.  Improve the matching speed of capturing brackets.
1164    
1165    10. Changed the meaning of \X so that it now matches a Unicode extended
1166        grapheme cluster.
1167    
1168    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1169        out POSIX threads when JIT support is configured.
1170    
1171    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1172    
1173    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1174        structure could go wrong in environments where size_t is not the same size
1175        as int.
1176    
1177    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1178        set.
1179    
1180    15. The EBCDIC support had decayed; later updates to the code had included
1181        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1182        general tidy up of EBCDIC-related issues, and the documentation was also
1183        not quite right. There is now a test that can be run on ASCII systems to
1184        check some of the EBCDIC-related things (but is it not a full test).
1185    
1186    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1187        in a small tidy to the code.
1188    
1189    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1190    
1191    18. If the --only-matching (-o) option in pcregrep is specified multiple
1192        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1193        substrings matched by the 1st and 2nd capturing parentheses. A separating
1194        string can be specified by --om-separator (default empty).
1195    
1196    19. Improving the first n character searches.
1197    
1198    20. Turn case lists for horizontal and vertical white space into macros so that
1199        they are defined only once.
1200    
1201    21. This set of changes together give more compatible Unicode case-folding
1202        behaviour for characters that have more than one other case when UCP
1203        support is available.
1204    
1205        (a) The Unicode property table now has offsets into a new table of sets of
1206            three or more characters that are case-equivalent. The MultiStage2.py
1207            script that generates these tables (the pcre_ucd.c file) now scans
1208            CaseFolding.txt instead of UnicodeData.txt for character case
1209            information.
1210    
1211        (b) The code for adding characters or ranges of characters to a character
1212            class has been abstracted into a generalized function that also handles
1213            case-independence. In UTF-mode with UCP support, this uses the new data
1214            to handle characters with more than one other case.
1215    
1216        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1217            whose other case is greater than 256 are now correctly matched
1218            caselessly. Previously, the high codepoint matched the low one, but not
1219            vice versa.
1220    
1221        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1222            of the new class addition function, using character lists defined as
1223            macros alongside the case definitions of 20 above.
1224    
1225        (e) Caseless back references now work with characters that have more than
1226            one other case.
1227    
1228        (f) General caseless matching of characters with more than one other case
1229            is supported.
1230    
1231    22. Unicode character properties were updated from Unicode 6.2.0
1232    
1233    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1234    
1235    24. Add support for 32-bit character strings, and UTF-32
1236    
1237    25. Major JIT compiler update (code refactoring and bugfixing).
1238        Experimental Sparc 32 support is added.
1239    
1240    26. Applied a modified version of Daniel Richard G's patch to create
1241        pcre.h.generic and config.h.generic by "make" instead of in the
1242        PrepareRelease script.
1243    
1244    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1245        pcre_compile.c when checking for a zero character.
1246    
1247    28. Introducing a native interface for JIT. Through this interface, the compiled
1248        machine code can be directly executed. The purpose of this interface is to
1249        provide fast pattern matching, so several sanity checks are not performed.
1250        However, feature tests are still performed. The new interface provides
1251        1.4x speedup compared to the old one.
1252    
1253    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1254        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1255        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1256    
1257    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1258        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1259        the "old" RFC 2279). Instead, it ended up passing a negative length to
1260        pcre_exec().
1261    
1262    31. Add support for GCC's visibility feature to hide internal functions.
1263    
1264    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1265        "unknown -C option" after outputting 0 or 1.
1266    
1267    33. There is now support for generating a code coverage report for the test
1268        suite in environments where gcc is the compiler and lcov is installed. This
1269        is mainly for the benefit of the developers.
1270    
1271    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1272        unaddressable using valgrind annotations, allowing valgrind to detect
1273        invalid memory accesses. This is mainly for the benefit of the developers.
1274    
1275    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1276    
1277    26. Give configure error if --enable-cpp but no C++ compiler found.
1278    
1279    
1280    Version 8.31 06-July-2012
1281    -------------------------
1282    
1283    1.  Fixing a wrong JIT test case and some compiler warnings.
1284    
1285    2.  Removed a bashism from the RunTest script.
1286    
1287    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1288        to unsigned type, result still unsigned" that was given by an MS compiler
1289        on encountering the code "-sizeof(xxx)".
1290    
1291    4.  Partial matching support is added to the JIT compiler.
1292    
1293    5.  Fixed several bugs concerned with partial matching of items that consist
1294        of more than one character:
1295    
1296        (a) /^(..)\1/ did not partially match "aba" because checking references was
1297            done on an "all or nothing" basis. This also applied to repeated
1298            references.
1299    
1300        (b) \R did not give a hard partial match if \r was found at the end of the
1301            subject.
1302    
1303        (c) \X did not give a hard partial match after matching one or more
1304            characters at the end of the subject.
1305    
1306        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1307            a partial match for the string "\r".
1308    
1309        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1310            a partial match for a CR character at the end of the subject string.
1311    
1312    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1313        running pcretest, the text "(JIT)" added to the output whenever JIT is
1314        actually used to run the match.
1315    
1316    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1317        or /S+[+] with a digit between 1 and 7.
1318    
1319    8.  OP_NOT now supports any UTF character not just single-byte ones.
1320    
1321    9.  (*MARK) control verb is now supported by the JIT compiler.
1322    
1323    10. The command "./RunTest list" lists the available tests without actually
1324        running any of them. (Because I keep forgetting what they all are.)
1325    
1326    11. Add PCRE_INFO_MAXLOOKBEHIND.
1327    
1328    12. Applied a (slightly modified) user-supplied patch that improves performance
1329        when the heap is used for recursion (compiled with --disable-stack-for-
1330        recursion). Instead of malloc and free for each heap frame each time a
1331        logical recursion happens, frames are retained on a chain and re-used where
1332        possible. This sometimes gives as much as 30% improvement.
1333    
1334    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1335        call.
1336    
1337    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1338    
1339    15. It is now possible to link pcretest with libedit as an alternative to
1340        libreadline.
1341    
1342    16. (*COMMIT) control verb is now supported by the JIT compiler.
1343    
1344    17. The Unicode data tables have been updated to Unicode 6.1.0.
1345    
1346    18. Added --file-list option to pcregrep.
1347    
1348    19. Added binary file support to pcregrep, including the -a, --binary-files,
1349        -I, and --text options.
1350    
1351    20. The madvise function is renamed for posix_madvise for QNX compatibility
1352        reasons. Fixed by Giuseppe D'Angelo.
1353    
1354    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1355    
1356    22. Changed the option for creating symbolic links for 16-bit man pages from
1357        -s to -sf so that re-installing does not cause issues.
1358    
1359    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1360    
1361    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1362        matches in certain environments (the workspace was not being correctly
1363        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1364        some of the workspace data at the beginning of a restart.
1365    
1366    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1367        was not doing so when it should - probably a typo introduced by SVN 528
1368        (change 8.10/14).
1369    
1370    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1371        \w+ when the character tables indicated that \x{c4} was a word character.
1372        There were several related cases, all because the tests for doing a table
1373        lookup were testing for characters less than 127 instead of 255.
1374    
1375    27. If a pattern contains capturing parentheses that are not used in a match,
1376        their slots in the ovector are set to -1. For those that are higher than
1377        any matched groups, this happens at the end of processing. In the case when
1378        there were back references that the ovector was too small to contain
1379        (causing temporary malloc'd memory to be used during matching), and the
1380        highest capturing number was not used, memory off the end of the ovector
1381        was incorrectly being set to -1. (It was using the size of the temporary
1382        memory instead of the true size.)
1383    
1384    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1385        ovector size, it uses memory at the end of the block that it has got.
1386    
1387    29. Check for an overlong MARK name and give an error at compile time. The
1388        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1389    
1390    30. JIT compiler update.
1391    
1392    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1393        Rill for the patch.
1394    
1395    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1396    
1397    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1398    
1399    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1400        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1401        was enabled.
1402    
1403    35. Improve JIT code generation for greedy plus quantifier.
1404    
1405    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1406        1 to "aa" instead of to an empty string. The bug affected repeated groups
1407        that could potentially match an empty string.
1408    
1409    37. Optimizing single character iterators in JIT.
1410    
1411    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1412        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1413        codepoints that are too big for the mode are faulted, and in a UTF mode,
1414        disallowed codepoints are also faulted.
1415    
1416    39. If PCRE was compiled with UTF support, in three places in the DFA
1417        matcher there was code that should only have been obeyed in UTF mode, but
1418        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1419        processing when bytes with values greater than 127 were present. In 16-bit
1420        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1421        both cases the values are those that cannot be the first data item in a UTF
1422        character. The three items that might have provoked this were recursions,
1423        possessively repeated groups, and atomic groups.
1424    
1425    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1426        and pcregrep, because some OS require shared objects to be explicitly
1427        passed to ld, causing the link step to fail if they are not.
1428    
1429    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1430        mode, patterns that started with \h* or \R* might be incorrectly matched.
1431    
1432    
1433    Version 8.30 04-February-2012
1434    -----------------------------
1435    
1436    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1437        name is defined in ctype.h.
1438    
1439    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1440        only in quite long subpatterns.
1441    
1442    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1443        since it was replaced by pcre_fullinfo() in February 2000.
1444    
1445    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1446        match a (*MARK), and the match failed at the start of the subject, a
1447        reference to memory before the start of the subject could occur. This bug
1448        was introduced by fix 17 of release 8.21.
1449    
1450    5.  A reference to an unset group with zero minimum repetition was giving
1451        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1452        /(another)?(\1?)test/ matched against "hello world test". This bug was
1453        introduced in release 8.13.
1454    
1455    6.  Add support for 16-bit character strings (a large amount of work involving
1456        many changes and refactorings).
1457    
1458    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1459        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1460        from a file.
1461    
1462    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1463        rounding is not applied in this particular case).
1464    
1465    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1466        if they appear, or are escaped, in patterns.
1467    
1468    10. Get rid of a number of -Wunused-but-set-variable warnings.
1469    
1470    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1471        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1472        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1473        also returns the mark "x". This bug applied to capturing parentheses,
1474        non-capturing parentheses, and atomic parentheses. It also applied to some
1475        assertions.
1476    
1477    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1478        information out of configure.ac instead of relying on pcre.h.generic, which
1479        is not stored in the repository.
1480    
1481    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1482        -lreadline.
1483    
1484    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1485    
1486    15. Applied Graycode's patch to put the top-level frame on the stack rather
1487        than the heap when not using the stack for recursion. This gives a
1488        performance improvement in many cases when recursion is not deep.
1489    
1490    16. Experimental code added to "pcretest -C" to output the stack frame size.
1491    
1492    
1493    Version 8.21 12-Dec-2011
1494    ------------------------
1495    
1496    1.  Updating the JIT compiler.
1497    
1498    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1499        are added as well.
1500    
1501    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1502        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1503        calling _pcre_jit_exec. Some extra comments are added.
1504    
1505    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1506        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1507        was introduced by change 18 for 8.20.
1508    
1509    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1510        ECMA-262 standard.
1511    
1512    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1513        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1514        This bug was probably introduced by change 9 of 8.13.
1515    
1516    7.  While fixing 6 above, I noticed that a number of other items were being
1517        incorrectly rejected as "not fixed length". This arose partly because newer
1518        opcodes had not been added to the fixed-length checking code. I have (a)
1519        corrected the bug and added tests for these items, and (b) arranged for an
1520        error to occur if an unknown opcode is encountered while checking for fixed
1521        length instead of just assuming "not fixed length". The items that were
1522        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1523        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1524        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1525    
1526    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1527        being incorrectly compiled and would have given unpredicatble results.
1528    
1529    9.  A possessively repeated subpattern with minimum repeat count greater than
1530        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1531        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1532        the first (A) could occur when it should not.
1533    
1534    10. Add a cast and remove a redundant test from the code.
1535    
1536    11. JIT should use pcre_malloc/pcre_free for allocation.
1537    
1538    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1539        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1540        is anything other than /usr, -L is still shown).
1541    
1542    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1543    
1544    14. Perl does not support \N without a following name in a [] class; PCRE now
1545        also gives an error.
1546    
1547    15. If a forward reference was repeated with an upper limit of around 2000,
1548        it caused the error "internal error: overran compiling workspace". The
1549        maximum number of forward references (including repeats) was limited by the
1550        internal workspace, and dependent on the LINK_SIZE. The code has been
1551        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1552        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1553        of around 200,000 forward references. While doing this, I also speeded up
1554        the filling in of repeated forward references.
1555    
1556    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1557        incorrectly expecting the subject to contain another "a" after the start.
1558    
1559    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1560        in the match, the SKIP should be ignored. This was not happening; instead
1561        the SKIP was being treated as NOMATCH. For patterns such as
1562        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1563        tested.
1564    
1565    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1566        now much more compatible with Perl, in particular in cases where the result
1567        is a non-match for a non-anchored pattern. For example, if
1568        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1569        "m", where previously it did not return a name. A side effect of this
1570        change is that for partial matches, the last encountered mark name is
1571        returned, as for non matches. A number of tests that were previously not
1572        Perl-compatible have been moved into the Perl-compatible test files. The
1573        refactoring has had the pleasing side effect of removing one argument from
1574        the match() function, thus reducing its stack requirements.
1575    
1576    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1577        subsequent uses of /S (without +) incorrectly behaved like /S+.
1578    
1579    21. Retrieve executable code size support for the JIT compiler and fixing
1580        some warnings.
1581    
1582    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1583        not work when the shorter character appeared right at the end of the
1584        subject string.
1585    
1586    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1587        systems.
1588    
1589    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1590        output it when the /M option is used in pcretest.
1591    
1592    25. The CheckMan script was not being included in the distribution. Also, added
1593        an explicit "perl" to run Perl scripts from the PrepareRelease script
1594        because this is reportedly needed in Windows.
1595    
1596    26. If study data was being save in a file and studying had not found a set of
1597        "starts with" bytes for the pattern, the data written to the file (though
1598        never used) was taken from uninitialized memory and so caused valgrind to
1599        complain.
1600    
1601    27. Updated RunTest.bat as provided by Sheri Pierce.
1602    
1603    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1604    
1605    29. Computation of memory usage for the table of capturing group names was
1606        giving an unnecessarily large value.
1607    
1608    
1609    Version 8.20 21-Oct-2011
1610    ------------------------
1611    
1612    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1613        a POSIX class. After further experiments with Perl, which convinced me that
1614        Perl has bugs and confusions, a closing square bracket is no longer allowed
1615        in a POSIX name. This bug also affected patterns with classes that started
1616        with full stops.
1617    
1618    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1619        captured substring, but while checking the failing first alternative,
1620        substring 1 is temporarily captured. If the output vector supplied to
1621        pcre_exec() was not big enough for this capture, the yield of the function
1622        was still zero ("insufficient space for captured substrings"). This cannot
1623        be totally fixed without adding another stack variable, which seems a lot
1624        of expense for a edge case. However, I have improved the situation in cases
1625        such as /(a)(b)x|abc/ matched against "abc", where the return code
1626        indicates that fewer than the maximum number of slots in the ovector have
1627        been set.
1628    
1629    3.  Related to (2) above: when there are more back references in a pattern than
1630        slots in the output vector, pcre_exec() uses temporary memory during
1631        matching, and copies in the captures as far as possible afterwards. It was
1632        using the entire output vector, but this conflicts with the specification
1633        that only 2/3 is used for passing back captured substrings. Now it uses
1634        only the first 2/3, for compatibility. This is, of course, another edge
1635        case.
1636    
1637    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1638        main code base, and can be used by building with --enable-jit. When this is
1639        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1640        runtime --no-jit option is given.
1641    
1642    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1643        ovector, the return from the function was zero, implying that there were
1644        other matches that did not fit. The correct "exactly full" value is now
1645        returned.
1646    
1647    6.  If a subpattern that was called recursively or as a subroutine contained
1648        (*PRUNE) or any other control that caused it to give a non-standard return,
1649        invalid errors such as "Error -26 (nested recursion at the same subject
1650        position)" or even infinite loops could occur.
1651    
1652    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1653        computing the minimum length on reaching *ACCEPT, and so ended up with the
1654        wrong value of 1 rather than 0. Further investigation indicates that
1655        computing a minimum subject length in the presence of *ACCEPT is difficult
1656        (think back references, subroutine calls), and so I have changed the code
1657        so that no minimum is registered for a pattern that contains *ACCEPT.
1658    
1659    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1660        it was not handled as intended. [But see 16 below.]
1661    
1662    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1663        Sheri Pierce.
1664    
1665    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1666        the first byte in a match must be "a".
1667    
1668    11. Change 17 for 8.13 increased the recursion depth for patterns like
1669        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1670        pattern contains any instances of (*THEN). If it does not, the old
1671        optimizations are restored. It would be nice to do this on a per-group
1672        basis, but at the moment that is not feasible.
1673    
1674    12. In some environments, the output of pcretest -C is CRLF terminated. This
1675        broke RunTest's code that checks for the link size. A single white space
1676        character after the value is now allowed for.
1677    
1678    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1679        For "fr", it uses the Windows-specific input and output files.
1680    
1681    14. If (*THEN) appeared in a group that was called recursively or as a
1682        subroutine, it did not work as intended. [But see next item.]
1683    
1684    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1685        pattern fragments (but not containing any | characters). If A and B are
1686        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1687        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1688        D. In other words, Perl considers parentheses that do not contain any |
1689        characters to be part of a surrounding alternative, whereas PCRE was
1690        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1691        differently. PCRE now behaves in the same way as Perl, except in the case
1692        of subroutine/recursion calls such as (?1) which have in any case always
1693        been different (but PCRE had them first :-).
1694    
1695    16. Related to 15 above: Perl does not treat the | in a conditional group as
1696        creating alternatives. Such a group is treated in the same way as an
1697        ordinary group without any | characters when processing (*THEN). PCRE has
1698        been changed to match Perl's behaviour.
1699    
1700    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1701        RunGrepTest script failed.
1702    
1703    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1704        inevitable for groups that contain captures, but it can lead to a lot of
1705        stack use in large patterns. The old behaviour has been restored for atomic
1706        groups that do not contain any capturing parentheses.
1707    
1708    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1709        suppress the check for a minimum subject length at run time. (If it was
1710        given to pcre_exec() or pcre_dfa_exec() it did work.)
1711    
1712    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1713        fail to work when decoding hex characters in data strings in EBCDIC
1714        environments.
1715    
1716    21. It appears that in at least one Mac OS environment, the isxdigit() function
1717        is implemented as a macro that evaluates to its argument more than once,
1718        contravening the C 90 Standard (I haven't checked a later standard). There
1719        was an instance in pcretest which caused it to go wrong when processing
1720        \x{...} escapes in subject strings. The has been rewritten to avoid using
1721        things like p++ in the argument of isxdigit().
1722    
1723    
1724    Version 8.13 16-Aug-2011
1725    ------------------------
1726    
1727    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1728    
1729    2.  Two minor typos in pcre_internal.h have been fixed.
1730    
1731    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1732        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1733        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1734    
1735    4.  There were a number of related bugs in the code for matching backrefences
1736        caselessly in UTF-8 mode when codes for the characters concerned were
1737        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1738        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1739        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1740        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1741        2-byte code at the end of the subject (it thought there wasn't enough data
1742        left).
1743    
1744    5.  Comprehensive information about what went wrong is now returned by
1745        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1746        as the output vector has at least 2 elements. The offset of the start of
1747        the failing character and a reason code are placed in the vector.
1748    
1749    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1750        now returned is for the first byte of the failing character, instead of the
1751        last byte inspected. This is an incompatible change, but I hope it is small
1752        enough not to be a problem. It makes the returned offset consistent with
1753        pcre_exec() and pcre_dfa_exec().
1754    
1755    7.  pcretest now gives a text phrase as well as the error number when
1756        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1757        failure, the offset and reason code are output.
1758    
1759    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1760        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1761        back over a single character (\n). This seems wrong (because it treated the
1762        two characters as a single entity when going forwards), conflicts with the
1763        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1764        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1765        has been changed.
1766    
1767    9.  Some internal refactoring has changed the processing so that the handling
1768        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1769        time (the PCRE_DOTALL option was changed this way some time ago: version
1770        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1771        which was always a bit of a fudge. It also means that there is one less
1772        argument for the match() function, which reduces its stack requirements
1773        slightly. This change also fixes an incompatibility with Perl: the pattern
1774        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1775    
1776    10. More internal refactoring has drastically reduced the number of recursive
1777        calls to match() for possessively repeated groups such as (abc)++ when
1778        using pcre_exec().
1779    
1780    11. While implementing 10, a number of bugs in the handling of groups were
1781        discovered and fixed:
1782    
1783        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1784        (a|)*(?1) gave a compile-time internal error.
1785        ((a|)+)+  did not notice that the outer group could match an empty string.
1786        (^a|^)+   was not marked as anchored.
1787        (.*a|.*)+ was not marked as matching at start or after a newline.
1788    
1789    12. Yet more internal refactoring has removed another argument from the match()
1790        function. Special calls to this function are now indicated by setting a
1791        value in a variable in the "match data" data block.
1792    
1793    13. Be more explicit in pcre_study() instead of relying on "default" for
1794        opcodes that mean there is no starting character; this means that when new
1795        ones are added and accidentally left out of pcre_study(), testing should
1796        pick them up.
1797    
1798    14. The -s option of pcretest has been documented for ages as being an old
1799        synonym of -m (show memory usage). I have changed it to mean "force study
1800        for every regex", that is, assume /S for every regex. This is similar to -i
1801        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1802        using it. It makes it easier to run collections of tests with and without
1803        study enabled, and thereby test pcre_study() more easily. All the standard
1804        tests are now run with and without -s (but some patterns can be marked as
1805        "never study" - see 20 below).
1806    
1807    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1808        restoration of the capturing data to the outer values was not happening
1809        correctly.
1810    
1811    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1812        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1813        pattern had matched an empty string, and so incorrectly returned a no
1814        match.
1815    
1816    17. There was optimizing code for the last branch of non-capturing parentheses,
1817        and also for the obeyed branch of a conditional subexpression, which used
1818        tail recursion to cut down on stack usage. Unfortunately, now that there is
1819        the possibility of (*THEN) occurring in these branches, tail recursion is
1820        no longer possible because the return has to be checked for (*THEN). These
1821        two optimizations have therefore been removed. [But see 8.20/11 above.]
1822    
1823    18. If a pattern containing \R was studied, it was assumed that \R always
1824        matched two bytes, thus causing the minimum subject length to be
1825        incorrectly computed because \R can also match just one byte.
1826    
1827    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1828        was incorrectly computed.
1829    
1830    20. If /S is present twice on a test pattern in pcretest input, it now
1831        *disables* studying, thereby overriding the use of -s on the command line
1832        (see 14 above). This is necessary for one or two tests to keep the output
1833        identical in both cases.
1834    
1835    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1836        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1837    
1838    22. When an atomic group that contained a capturing parenthesis was
1839        successfully matched, but the branch in which it appeared failed, the
1840        capturing was not being forgotten if a higher numbered group was later
1841        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1842        group 1 to "a", when in fact it should be unset. This applied to multi-
1843        branched capturing and non-capturing groups, repeated or not, and also to
1844        positive assertions (capturing in negative assertions does not happen
1845        in PCRE) and also to nested atomic groups.
1846    
1847    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1848        subject after a captured substring, to make it easier to tell which of a
1849        number of identical substrings has been captured.
1850    
1851    24. The way atomic groups are processed by pcre_exec() has been changed so that
1852        if they are repeated, backtracking one repetition now resets captured
1853        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1854        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1855        "aaa". Previously, it would have been "a". As part of this code
1856        refactoring, the way recursive calls are handled has also been changed.
1857    
1858    25. If an assertion condition captured any substrings, they were not passed
1859        back unless some other capturing happened later. For example, if
1860        (?(?=(a))a) was matched against "a", no capturing was returned.
1861    
1862    26. When studying a pattern that contained subroutine calls or assertions,
1863        the code for finding the minimum length of a possible match was handling
1864        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1865        group 1 called group 2 while simultaneously a separate group 2 called group
1866        1). A stack overflow occurred in this case. I have fixed this by limiting
1867        the recursion depth to 10.
1868    
1869    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1870        Fortmann. This supports explicit test numbers on the command line, and has
1871        argument validation and error reporting.
1872    
1873    28. An instance of \X with an unlimited repeat could fail if at any point the
1874        first character it looked at was a mark character.
1875    
1876    29. Some minor code refactoring concerning Unicode properties and scripts
1877        should reduce the stack requirement of match() slightly.
1878    
1879    30. Added the '=' option to pcretest to check the setting of unused capturing
1880        slots at the end of the pattern, which are documented as being -1, but are
1881        not included in the return count.
1882    
1883    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1884        compiled something random. Now it gives a compile-time error (as does
1885        Perl).
1886    
1887    32. A *MARK encountered during the processing of a positive assertion is now
1888        recorded and passed back (compatible with Perl).
1889    
1890    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1891        had alternative anchored branches, the search for a second match in a line
1892        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1893        matched the line "0102" twice. The same bug affected patterns that started
1894        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1895        twice.
1896    
1897    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1898        does, and because of capturing effects, quantifying parenthesized
1899        assertions may at times be useful. Quantifiers are now allowed for
1900        parenthesized assertions.
1901    
1902    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1903    
1904    36. \g was being checked for fancy things in a character class, when it should
1905        just be a literal "g".
1906    
1907    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1908        appearance of a nested POSIX class supersedes an apparent external class.
1909        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1910        unescaped square brackets may also appear as part of class names. For
1911        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1912        more like Perl. (But see 8.20/1 above.)
1913    
1914    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1915        was because it thought it was \N{name}, which is not supported).
1916    
1917    39. Add minix to OS list not supporting the -S option in pcretest.
1918    
1919    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1920        cannot analyze patterns in sufficient detail to catch mutual recursions
1921        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1922        subgroup is called recursively as a subpattern for a second time at the
1923        same position in the subject string. In previous releases this might have
1924        been caught by the recursion limit, or it might have run out of stack.
1925    
1926    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1927        happen only once. PCRE was, however incorrectly giving a compile time error
1928        "recursive call could loop indefinitely" because it cannot analyze the
1929        pattern in sufficient detail. The compile time test no longer happens when
1930        PCRE is compiling a conditional subpattern, but actual runaway loops are
1931        now caught at runtime (see 40 above).
1932    
1933    42. It seems that Perl allows any characters other than a closing parenthesis
1934        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1935        has been changed to be the same.
1936    
1937    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1938        as not to get warnings when autogen.sh is called. Also changed
1939        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1940    
1941    44. To help people who use pcregrep to scan files containing exceedingly long
1942        lines, the following changes have been made:
1943    
1944        (a) The default value of the buffer size parameter has been increased from
1945            8K to 20K. (The actual buffer used is three times this size.)
1946    
1947        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1948            PCRE is built.
1949    
1950        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1951            to be set at run time.
1952    
1953        (d) Numerical values in pcregrep options can be followed by K or M, for
1954            example --buffer-size=50K.
1955    
1956        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1957            given and the return code is set to 2.
1958    
1959    45. Add a pointer to the latest mark to the callout data block.
1960    
1961    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1962        partial match of an empty string instead of no match. This was specific to
1963        the use of ".".
1964    
1965    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1966        complete match instead of a partial match. This bug was dependent on both
1967        the PCRE_UTF8 and PCRE_DOTALL options being set.
1968    
1969    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1970        starting byte set, because \b was not being ignored.
1971    
1972    
1973    Version 8.12 15-Jan-2011
1974    ------------------------
1975    
1976    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1977        checks for such things as part of the documentation building process.
1978    
1979    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1980        --match-limit and --recursion-limit options (added for 8.11). In
1981        particular, this made one of the standard tests fail. (The integer value
1982        went into the wrong half of a long int.)
1983    
1984    3.  If the --colour option was given to pcregrep with -v (invert match), it
1985        did strange things, either producing crazy output, or crashing. It should,
1986        of course, ignore a request for colour when reporting lines that do not
1987        match.
1988    
1989    4.  Another pcregrep bug caused similar problems if --colour was specified with
1990        -M (multiline) and the pattern match finished with a line ending.
1991    
1992    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1993        matched in multiline mode, the following line was shown as part of the
1994        match. This seems wrong, so I have changed it.
1995    
1996    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1997        the check for further matches in the same line (so they could be coloured)
1998        to overrun the end of the current line. If another match was found, it was
1999        incorrectly shown (and then shown again when found in the next line).
2000    
2001    7.  If pcregrep was compiled under Windows, there was a reference to the
2002        function pcregrep_exit() before it was defined. I am assuming this was
2003        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
2004        reported by a user. I've moved the definition above the reference.
2005    
2006    
2007    Version 8.11 10-Dec-2010
2008    ------------------------
2009    
2010    1.  (*THEN) was not working properly if there were untried alternatives prior
2011        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
2012        backtracked to try for "b" instead of moving to the next alternative branch
2013        at the same level (in this case, to look for "c"). The Perl documentation
2014        is clear that when (*THEN) is backtracked onto, it goes to the "next
2015        alternative in the innermost enclosing group".
2016    
2017    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
2018        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
2019        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
2020        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
2021        (*THEN).
2022    
2023    3.  If \s appeared in a character class, it removed the VT character from
2024        the class, even if it had been included by some previous item, for example
2025        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
2026        of \s, but is part of the POSIX "space" class.)
2027    
2028    4.  A partial match never returns an empty string (because you can always
2029        match an empty string at the end of the subject); however the checking for
2030        an empty string was starting at the "start of match" point. This has been
2031        changed to the "earliest inspected character" point, because the returned
2032        data for a partial match starts at this character. This means that, for
2033        example, /(?<=abc)def/ gives a partial match for the subject "abc"
2034        (previously it gave "no match").
2035    
2036    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
2037        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
2038        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
2039        has an implication that the given string is incomplete (because a partial
2040        match is preferred over a full match). For this reason, these items now
2041        give a partial match in this situation. [Aside: previously, the one case
2042        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
2043        match rather than a full match, which was wrong by the old rules, but is
2044        now correct.]
2045    
2046    6.  There was a bug in the handling of #-introduced comments, recognized when
2047        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
2048        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
2049        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
2050        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
2051        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
2052        places in pcre_compile().
2053    
2054    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
2055        comments when looking ahead for named forward references to subpatterns,
2056        the only newline sequence it recognized was NL. It now handles newlines
2057        according to the set newline convention.
2058    
2059    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
2060        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
2061        cater for a lack of strerror(). These oversights have been fixed.
2062    
2063    9.  Added --match-limit and --recursion-limit to pcregrep.
2064    
2065    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
2066    
2067    11. When the -o option was used, pcregrep was setting a return code of 1, even
2068        when matches were found, and --line-buffered was not being honoured.
2069    
2070    12. Added an optional parentheses number to the -o and --only-matching options
2071        of pcregrep.
2072    
2073    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
2074        can match an empty string. The code to do it in pcretest and pcredemo
2075        needed fixing:
2076    
2077        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
2078            only one byte after an empty string match just before CRLF (this case
2079            just got forgotten; "any" and "anycrlf" were OK).
2080    
2081        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
2082            mode when an empty string match preceded an ASCII character followed by
2083            a non-ASCII character. (The code for advancing by one character rather
2084            than one byte was nonsense.)
2085    
2086        (c) The pcredemo.c sample program did not have any code at all to handle
2087            the cases when CRLF is a valid newline sequence.
2088    
2089    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2090        as a starting offset was within the subject string. There is now a new
2091        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2092        negative or greater than the length of the string. In order to test this,
2093        pcretest is extended to allow the setting of negative starting offsets.
2094    
2095    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2096        starting offset points to the beginning of a UTF-8 character was
2097        unnecessarily clumsy. I tidied it up.
2098    
2099    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2100        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2101    
2102    17. Nobody had reported that the --include_dir option, which was added in
2103        release 7.7 should have been called --include-dir (hyphen, not underscore)
2104        for compatibility with GNU grep. I have changed it to --include-dir, but
2105        left --include_dir as an undocumented synonym, and the same for
2106        --exclude-dir, though that is not available in GNU grep, at least as of
2107        release 2.5.4.
2108    
2109    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2110        characters from a string of bytes) have been redefined so as not to use
2111        loops, in order to improve performance in some environments. At the same
2112        time, I abstracted some of the common code into auxiliary macros to save
2113        repetition (this should not affect the compiled code).
2114    
2115    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2116        compile-time error is now given if \c is not followed by an ASCII
2117        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2118        different, and any byte value is allowed.)
2119    
2120    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2121        START_OPTIMIZE option, which is now allowed at compile time - but just
2122        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2123        to pcregrep and other applications that have no direct access to PCRE
2124        options. The new /Y option in pcretest sets this option when calling
2125        pcre_compile().
2126    
2127    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2128        back references. Groups containing recursive back references were forced to
2129        be atomic by that change, but in the case of named groups, the amount of
2130        memory required was incorrectly computed, leading to "Failed: internal
2131        error: code overflow". This has been fixed.
2132    
2133    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2134        pcretest.c, to avoid build problems in some Borland environments.
2135    
2136    
2137    Version 8.10 25-Jun-2010
2138    ------------------------
2139    
2140    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2141        THEN.
2142    
2143    2.  (*ACCEPT) was not working when inside an atomic group.
2144    
2145    3.  Inside a character class, \B is treated as a literal by default, but
2146        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2147        causes the error). The code is unchanged, but I tidied the documentation.
2148    
2149    4.  Inside a character class, PCRE always treated \R and \X as literals,
2150        whereas Perl faults them if its -w option is set. I have changed PCRE so
2151        that it faults them when PCRE_EXTRA is set.
2152    
2153    5.  Added support for \N, which always matches any character other than
2154        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2155    
2156    6.  When compiling pcregrep with newer versions of gcc which may have
2157        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2158        declared with attribute warn_unused_result" were given. Just casting the
2159        result to (void) does not stop the warnings; a more elaborate fudge is
2160        needed. I've used a macro to implement this.
2161    
2162    7.  Minor change to pcretest.c to avoid a compiler warning.
2163    
2164    8.  Added four artifical Unicode properties to help with an option to make
2165        \s etc use properties (see next item). The new properties are: Xan
2166        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2167    
2168    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2169        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2170        this option. Modified pcretest to add /W to test this facility. Added
2171        REG_UCP to make it available via the POSIX interface.
2172    
2173    10. Added --line-buffered to pcregrep.
2174    
2175    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2176        studied, and the match started with a letter with a code point greater than
2177        127 whose first byte was different to the first byte of the other case of
2178        the letter, the other case of this starting letter was not recognized
2179        (#976).
2180    
2181    12. If a pattern that was studied started with a repeated Unicode property
2182        test, for example, \p{Nd}+, there was the theoretical possibility of
2183        setting up an incorrect bitmap of starting bytes, but fortunately it could
2184        not have actually happened in practice until change 8 above was made (it
2185        added property types that matched character-matching opcodes).
2186    
2187    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2188        possible starting bytes for non-anchored patterns.
2189    
2190    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2191        \R, and also a number of cases that involve Unicode properties, both
2192        explicit and implicit when PCRE_UCP is set.
2193    
2194    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2195        input, it could crash or give wrong results if characters with values
2196        greater than 0xc0 were present in the subject string. (Detail: it assumed
2197        UTF-8 input when processing these items.)
2198    
2199    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2200        size_t is 64-bit (#991).
2201    
2202    17. Added a check for running out of memory when PCRE is compiled with
2203        --disable-stack-for-recursion (#990).
2204    
2205    18. If the last data line in a file for pcretest does not have a newline on
2206        the end, a newline was missing in the output.
2207    
2208    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2209        less than 128) in its various bitmaps. However, there is a facility for
2210        generating tables according to the current locale when PCRE is compiled. It
2211        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2212        characters, are recognized by isspace() and therefore were getting set in
2213        these tables, and indeed these tables seem to approximate to ISO 8859. This
2214        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2215        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2216        which of course cannot start UTF-8 characters. I have changed the code so
2217        that only real ASCII characters (less than 128) and the correct starting
2218        bytes for UTF-8 encodings are set for characters greater than 127 when in
2219        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2220        altogether.)
2221    
2222    20. Added the /T option to pcretest so as to be able to run tests with non-
2223        standard character tables, thus making it possible to include the tests
2224        used for 19 above in the standard set of tests.
2225    
2226    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2227        reference to a subpattern the other side of a comment that contains an
2228        opening parenthesis caused either an internal compiling error, or a
2229        reference to the wrong subpattern.
2230    
2231    
2232    Version 8.02 19-Mar-2010
2233    ------------------------
2234    
2235    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2236    
2237    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2238        configured.
2239    
2240    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2241        original author of that file, following a query about its status.
2242    
2243    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2244        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2245    
2246    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2247        quantifier applied to a forward-referencing subroutine call, could compile
2248        incorrect code or give the error "internal error: previously-checked
2249        referenced subpattern not found".
2250    
2251    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2252        variables to point to external functions. For these systems, therefore,
2253        pcre_malloc etc. are now initialized to local functions that call the
2254        relevant global functions.
2255    
2256    7.  There were two entries missing in the vectors called coptable and poptable
2257        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2258        I've fixed the data, and added a kludgy way of testing at compile time that
2259        the lengths are correct (equal to the number of opcodes).
2260    
2261    8.  Following on from 7, I added a similar kludge to check the length of the
2262        eint vector in pcreposix.c.
2263    
2264    9.  Error texts for pcre_compile() are held as one long string to avoid too
2265        much relocation at load time. To find a text, the string is searched,
2266        counting zeros. There was no check for running off the end of the string,
2267        which could happen if a new error number was added without updating the
2268        string.
2269    
2270    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2271    
2272    11. \K was not working if it appeared in an atomic group or in a group that
2273        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2274        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2275        the assertion is positive, but not if it is negative.
2276    
2277    12. Change 11 fortuitously reduced the size of the stack frame used in the
2278        "match()" function of pcre_exec.c by one pointer. Forthcoming
2279        implementation of support for (*MARK) will need an extra pointer on the
2280        stack; I have reserved it now, so that the stack frame size does not
2281        decrease.
2282    
2283    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2284        item in branch that calls a recursion is a subroutine call - as in the
2285        second branch in the above example - was incorrectly given the compile-
2286        time error "recursive call could loop indefinitely" because pcre_compile()
2287        was not correctly checking the subroutine for matching a non-empty string.
2288    
2289    14. The checks for overrunning compiling workspace could trigger after an
2290        overrun had occurred. This is a "should never occur" error, but it can be
2291        triggered by pathological patterns such as hundreds of nested parentheses.
2292        The checks now trigger 100 bytes before the end of the workspace.
2293    
2294    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2295    
2296    
2297    Version 8.01 19-Jan-2010
2298    ------------------------
2299    
2300    1.  If a pattern contained a conditional subpattern with only one branch (in
2301        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2302        computed the wrong minimum data length (which is of course zero for such
2303        subpatterns). This could cause incorrect "no match" results.
2304    
2305    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2306        the pattern is reset in the first branch, pcre_compile() failed with
2307        "internal error: code overflow at offset...". This happened only when
2308        the reset was to the original external option setting. (An optimization
2309        abstracts leading options settings into an external setting, which was the
2310        cause of this.)
2311    
2312    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2313        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2314        assertion pattern did not match (meaning that the assertion was true), it
2315        was incorrectly treated as false if the SKIP had been reached during the
2316        matching. This also applied to assertions used as conditions.
2317    
2318    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2319        assertion subpattern, including such a pattern used as a condition,
2320        unpredictable results occurred, instead of the error return
2321        PCRE_ERROR_DFA_UITEM.
2322    
2323    5.  The C++ GlobalReplace function was not working like Perl for the special
2324        situation when an empty string is matched. It now does the fancy magic
2325        stuff that is necessary.
2326    
2327    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2328        removed. (These were left over from very, very early versions of PCRE.)
2329    
2330    7.  Some cosmetic changes to the code to make life easier when compiling it
2331        as part of something else:
2332    
2333        (a) Change DEBUG to PCRE_DEBUG.
2334    
2335        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2336            called "current" as "current_branch", to prevent a collision with the
2337            Linux macro when compiled as a kernel module.
2338    
2339        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2340            prevent a collision with the Linux macro when compiled as a kernel
2341            module.
2342    
2343    8.  In pcre_compile() there are some checks for integer overflows that used to
2344        cast potentially large values to (double). This has been changed to that
2345        when building, a check for int64_t is made, and if it is found, it is used
2346        instead, thus avoiding the use of floating point arithmetic. (There is no
2347        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2348        double.
2349    
2350    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2351        2005 (difference between two addresses compared to an unsigned value).
2352    
2353    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2354        custom one, because of the following reported problem in Windows:
2355    
2356          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2357              under Win32.
2358          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2359              therefore missing the function definition.
2360          - The compiler thus generates a "C" signature for the test function.
2361          - The linker fails to find the "C" function.
2362          - PCRE fails to configure if asked to do so against libbz2.
2363    
2364    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2365        messages were output:
2366    
2367          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2368          rerunning libtoolize, to keep the correct libtool macros in-tree.
2369          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2370    
2371        I have done both of these things.
2372    
2373    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2374        most of the time, it *can* run out if it is given a pattern that contains a
2375        runaway infinite recursion. I updated the discussion in the pcrestack man
2376        page.
2377    
2378    13. Now that we have gone to the x.xx style of version numbers, the minor
2379        version may start with zero. Using 08 or 09 is a bad idea because users
2380        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2381        interpreted as invalid octal numbers. I've updated the previous comment in
2382        configure.ac, and also added a check that gives an error if 08 or 09 are
2383        used.
2384    
2385    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2386        causing partial matching to fail when the end of the subject matched \W
2387        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2388    
2389    15. There were some discrepancies between the declarations in pcre_internal.h
2390        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2391        their definitions. The declarations used "const uschar *" and the
2392        definitions used USPTR. Even though USPTR is normally defined as "const
2393        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2394        reported that: "This difference in casting confuses some C++ compilers, for
2395        example, SunCC recognizes above declarations as different functions and
2396        generates broken code for hbpcre." I have changed the declarations to use
2397        USPTR.
2398    
2399    16. GNU libtool is named differently on some systems. The autogen.sh script now
2400        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2401        (FreeBSD).
2402    
2403    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2404        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2405        comment: "Figure out how to create a longlong from a string: strtoll and
2406        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2407        instance, but it only takes 2 args instead of 3!"
2408    
2409    18. A subtle bug concerned with back references has been fixed by a change of
2410        specification, with a corresponding code fix. A pattern such as
2411        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2412        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2413        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2414        same bug. Such groups have to be quantified to be useful, or contained
2415        inside another quantified group. (If there's no repetition, the reference
2416        can never match.) The problem arises because, having left the group and
2417        moved on to the rest of the pattern, a later failure that backtracks into
2418        the group uses the captured value from the final iteration of the group
2419        rather than the correct earlier one. I have fixed this in PCRE by forcing
2420        any group that contains a reference to itself to be an atomic group; that
2421        is, there cannot be any backtracking into it once it has completed. This is
2422        similar to recursive and subroutine calls.
2423    
2424    
2425    Version 8.00 19-Oct-09
2426    ----------------------
2427    
2428    1.  The table for translating pcre_compile() error codes into POSIX error codes
2429        was out-of-date, and there was no check on the pcre_compile() error code
2430        being within the table. This could lead to an OK return being given in
2431        error.
2432    
2433    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2434        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2435        in a Windows environment.
2436    
2437    3.  The pcregrep --count option prints the count for each file even when it is
2438        zero, as does GNU grep. However, pcregrep was also printing all files when
2439        --files-with-matches was added. Now, when both options are given, it prints
2440        counts only for those files that have at least one match. (GNU grep just
2441        prints the file name in this circumstance, but including the count seems
2442        more useful - otherwise, why use --count?) Also ensured that the
2443        combination -clh just lists non-zero counts, with no names.
2444    
2445    4.  The long form of the pcregrep -F option was incorrectly implemented as
2446        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2447        but it seems right to fix it, and I didn't think it was worth preserving
2448        the old behaviour.
2449    
2450    5.  The command line items --regex=pattern and --regexp=pattern were not
2451        recognized by pcregrep, which required --regex pattern or --regexp pattern
2452        (with a space rather than an '='). The man page documented the '=' forms,
2453        which are compatible with GNU grep; these now work.
2454    
2455    6.  No libpcreposix.pc file was created for pkg-config; there was just
2456        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2457    
2458    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2459        when UCP support is not needed, by modifying the Python script that
2460        generates it from Unicode data files. This should not matter if the module
2461        is correctly used as a library, but I received one complaint about 50K of
2462        unwanted data. My guess is that the person linked everything into his
2463        program rather than using a library. Anyway, it does no harm.
2464    
2465    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2466        was a minimum greater than 1 for a wide character in a possessive
2467        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2468        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2469        character. Chaos in the form of incorrect output or a compiling loop could
2470        result.
2471    
2472    9.  The restrictions on what a pattern can contain when partial matching is
2473        requested for pcre_exec() have been removed. All patterns can now be
2474        partially matched by this function. In addition, if there are at least two
2475        slots in the offset vector, the offset of the earliest inspected character
2476        for the match and the offset of the end of the subject are set in them when
2477        PCRE_ERROR_PARTIAL is returned.
2478    
2479    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2480        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2481        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2482        and may be more useful for multi-segment matching.
2483    
2484    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2485        used to be given if ever the end of the subject was reached; now it is
2486        given only if matching could not proceed because another character was
2487        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2488        string "Z", which now yields "no match" instead of "partial match". In the
2489        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2490        final character ended with (*FAIL).
2491    
2492    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2493        if the pattern had a "must contain" character that was already found in the
2494        earlier partial match, unless partial matching was again requested. For
2495        example, with the pattern /dog.(body)?/, the "must contain" character is
2496        "g". If the first part-match was for the string "dog", restarting with
2497        "sbody" failed. This bug has been fixed.
2498    
2499    13. The string returned by pcre_dfa_exec() after a partial match has been
2500        changed so that it starts at the first inspected character rather than the
2501        first character of the match. This makes a difference only if the pattern
2502        starts with a lookbehind assertion or \b or \B (\K is not supported by
2503        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2504        matching functions compatible, and I think it's the right thing to do.
2505    
2506    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2507        so that the demonstration program is easily available in environments where
2508        PCRE has not been installed from source.
2509    
2510    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2511        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2512        library.
2513    
2514    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2515        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2516        is not the first non-POSIX option to be added. Clearly some people find
2517        these options useful.
2518    
2519    17. If a caller to the POSIX matching function regexec() passes a non-zero
2520        value for nmatch with a NULL value for pmatch, the value of
2521        nmatch is forced to zero.
2522    
2523    18. RunGrepTest did not have a test for the availability of the -u option of
2524        the diff command, as RunTest does. It now checks in the same way as
2525        RunTest, and also checks for the -b option.
2526    
2527    19. If an odd number of negated classes containing just a single character
2528        interposed, within parentheses, between a forward reference to a named
2529        subpattern and the definition of the subpattern, compilation crashed with
2530        an internal error, complaining that it could not find the referenced
2531        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2532        [The bug was that it was starting one character too far in when skipping
2533        over the character class, thus treating the ] as data rather than
2534        terminating the class. This meant it could skip too much.]
2535    
2536    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2537        /g option in pcretest when the pattern contains \K, which makes it possible
2538        to have an empty string match not at the start, even when the pattern is
2539        anchored. Updated pcretest and pcredemo to use this option.
2540    
2541    21. If the maximum number of capturing subpatterns in a recursion was greater
2542        than the maximum at the outer level, the higher number was returned, but
2543        with unset values at the outer level. The correct (outer level) value is
2544        now given.
2545    
2546    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2547        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2548        make it do so. The string so far is captured, making this feature
2549        compatible with Perl.
2550    
2551    23. The tests have been re-organized, adding tests 11 and 12, to make it
2552        possible to check the Perl 5.10 features against Perl 5.10.
2553    
2554    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2555        pattern matches a fixed length string. PCRE did not allow this; now it
2556        does. Neither allows recursion.
2557    
2558    25. I finally figured out how to implement a request to provide the minimum
2559        length of subject string that was needed in order to match a given pattern.
2560        (It was back references and recursion that I had previously got hung up
2561        on.) This code has now been added to pcre_study(); it finds a lower bound
2562        to the length of subject needed. It is not necessarily the greatest lower
2563        bound, but using it to avoid searching strings that are too short does give
2564        some useful speed-ups. The value is available to calling programs via
2565        pcre_fullinfo().
2566    
2567    26. While implementing 25, I discovered to my embarrassment that pcretest had
2568        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2569        study optimizations had never been tested with that matching function.
2570        Oops. What is worse, even when it was passed study data, there was a bug in
2571        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2572        were also very few tests of studied patterns with pcre_dfa_exec().
2573    
2574    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2575        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2576        on the other side of the coin, they are no longer allowed to have different
2577        names, because these cannot be distinguished in PCRE, and this has caused
2578        confusion. (This is a difference from Perl.)
2579    
2580    28. When duplicate subpattern names are present (necessarily with different
2581        numbers, as required by 27 above), and a test is made by name in a
2582        conditional pattern, either for a subpattern having been matched, or for
2583        recursion in such a pattern, all the associated numbered subpatterns are
2584        tested, and the overall condition is true if the condition is true for any
2585        one of them. This is the way Perl works, and is also more like the way
2586        testing by number works.
2587    
2588    
2589    Version 7.9 11-Apr-09
2590    ---------------------
2591    
2592    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2593        (pcretest), all targets were linked against these libraries. This included
2594        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2595        libraries. This caused unwanted dependencies to be created. This problem
2596        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2597        pcretest is linked with readline.
2598    
2599    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2600        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2601        moved outside it again, because FALSE and TRUE are already defined in AIX,
2602        but BOOL is not.
2603    
2604    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2605        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2606    
2607    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2608        hyphens) following file names and line numbers when outputting matching
2609        lines. This is not true; no spaces are inserted. I have also clarified the
2610        wording for the --colour (or --color) option.
2611    
2612    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2613        was not coloured; this is different to GNU grep, so I have changed it to be
2614        the same.
2615    
2616    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2617        each matching line was coloured. Now it goes on to look for further matches
2618        of any of the test patterns, which is the same behaviour as GNU grep.
2619    
2620    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2621        doesn't make sense to accept an empty string match in pcregrep, so I have
2622        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2623        seems to be how GNU grep behaves. [But see later change 40 for release
2624        8.33.]
2625    
2626    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2627        start or after a newline", because the conditional assertion was not being
2628        correctly handled. The rule now is that both the assertion and what follows
2629        in the first alternative must satisfy the test.
2630    
2631    9.  If auto-callout was enabled in a pattern with a conditional group whose
2632        condition was an assertion, PCRE could crash during matching, both with
2633        pcre_exec() and pcre_dfa_exec().
2634    
2635    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2636        used for matching.
2637    
2638    11. Unicode property support in character classes was not working for
2639        characters (bytes) greater than 127 when not in UTF-8 mode.
2640    
2641    12. Added the -M command line option to pcretest.
2642    
2643    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2644    
2645    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2646    
2647    16. Added comments and documentation about mis-use of no_arg in the C++
2648        wrapper.
2649    
2650    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2651        from Martin Jerabek that uses macro names for all relevant character and
2652        string constants.
2653    
2654    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2655        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2656        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2657        these, but not everybody uses configure.
2658    
2659    19. A conditional group that had only one branch was not being correctly
2660        recognized as an item that could match an empty string. This meant that an
2661        enclosing group might also not be so recognized, causing infinite looping
2662        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2663        with the subject "ab", where knowledge that the repeated group can match
2664        nothing is needed in order to break the loop.
2665    
2666    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2667        exec(), but without supplying a callout function, matching went wrong.
2668    
2669    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2670        leak if the size of the offset vector was greater than 30. When the vector
2671        is smaller, the saved offsets during recursion go onto a local stack
2672        vector, but for larger vectors malloc() is used. It was failing to free
2673        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2674        error, in fact).
2675    
2676    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2677        heapframe that is used only when UTF-8 support is enabled. This caused no
2678        problem, but was untidy.
2679    
2680    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2681        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2682        included within another project.
2683    
2684    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2685        slightly modified by me:
2686    
2687          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2688              not building pcregrep.
2689    
2690          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2691              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2692    
2693    25. Forward references, both numeric and by name, in patterns that made use of
2694        duplicate group numbers, could behave incorrectly or give incorrect errors,
2695        because when scanning forward to find the reference group, PCRE was not
2696        taking into account the duplicate group numbers. A pattern such as
2697        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2698    
2699    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2700        the feature of a custom pointer more persuasive (as requested by a user).
2701    
2702    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2703        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2704        pre-defined.
2705    
2706    28. Added support for (*UTF8) at the start of a pattern.
2707    
2708    29. Arrange for flags added by the "release type" setting in CMake to be shown
2709        in the configuration summary.
2710    
2711    
2712    Version 7.8 05-Sep-08
2713    ---------------------
2714    
2715    1.  Replaced UCP searching code with optimized version as implemented for Ad
2716        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2717        stage table and inline lookup instead of a function, giving speed ups of 2
2718        to 5 times on some simple patterns that I tested. Permission was given to
2719        distribute the MultiStage2.py script that generates the tables (it's not in
2720        the tarball, but is in the Subversion repository).
2721    
2722    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2723        scripts.
2724    
2725    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2726        a group with a zero qualifier. The result of the study could be incorrect,
2727        or the function might crash, depending on the pattern.
2728    
2729    4.  Caseless matching was not working for non-ASCII characters in back
2730        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2731        It now works when Unicode Property Support is available.
2732    
2733    5.  In pcretest, an escape such as \x{de} in the data was always generating
2734        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2735        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2736        truncation.
2737    
2738    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2739    
2740    7.  Added two (int) casts to pcregrep when printing the difference of two
2741        pointers, in case they are 64-bit values.
2742    
2743    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2744        test 2 if it fails.
2745    
2746    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2747        and a #define of that name to empty if it is not externally set. This is to
2748        allow users of MSVC to set it if necessary.
2749    
2750    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2751        the convenience functions in the pcre_get.c source file.
2752    
2753    11. An option change at the start of a pattern that had top-level alternatives
2754        could cause overwriting and/or a crash. This command provoked a crash in
2755        some environments:
2756    
2757          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2758    
2759        This potential security problem was recorded as CVE-2008-2371.
2760    
2761    12. For a pattern where the match had to start at the beginning or immediately
2762        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2763        pcre_dfa_exec() could read past the end of the passed subject if there was
2764        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2765        pcretest so that it places the subject at the end of its malloc-ed buffer.
2766    
2767    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2768        exec() might read past the end of the data buffer in UTF-8 mode.
2769    
2770    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2771        the data contained the byte 0x85 as part of a UTF-8 character within its
2772        first line. This applied both to normal and DFA matching.
2773    
2774    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2775        /^[^d]*?$/8 failed to match "abc".
2776    
2777    16. Added a missing copyright notice to pcrecpp_internal.h.
2778    
2779    17. Make it more clear in the documentation that values returned from
2780        pcre_exec() in ovector are byte offsets, not character counts.
2781    
2782    18. Tidied a few places to stop certain compilers from issuing warnings.
2783    
2784    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2785        supplied by Stefan Weber. I made a further small update for 7.8 because
2786        there is a change of source arrangements: the pcre_searchfuncs.c module is
2787        replaced by pcre_ucd.c.
2788    
2789    
2790    Version 7.7 07-May-08
2791    ---------------------
2792    
2793    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2794        a string to a long long, pretend we don't even have a long long." This is
2795        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2796    
2797    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2798        pre-7.6 versions, which defined a global no_arg variable instead of putting
2799        it in the RE class. (See also #8 below.)
2800    
2801    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2802        Lopes.
2803    
2804    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2805    
2806        (1) The include/exclude patterns were being applied to the whole pathnames
2807            of files, instead of just to the final components.
2808    
2809        (2) If there was more than one level of directory, the subdirectories were
2810            skipped unless they satisfied the include/exclude conditions. This is
2811            inconsistent with GNU grep (and could even be seen as contrary to the
2812            pcregrep specification - which I improved to make it absolutely clear).
2813            The action now is always to scan all levels of directory, and just
2814            apply the include/exclude patterns to regular files.
2815    
2816    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2817        --exclude_dir in the tests to avoid scanning .svn directories.
2818    
2819    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2820        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2821        doesn't support NULs in patterns.
2822    
2823    7.  Added some missing "const"s to declarations of static tables in
2824        pcre_compile.c and pcre_dfa_exec.c.
2825    
2826    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2827        caused by fix #2  above. (Subsequently also a second patch to fix the
2828        first patch. And a third patch - this was a messy problem.)
2829    
2830    9.  Applied Craig's patch to remove the use of push_back().
2831    
2832    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2833        matching function regexec().
2834    
2835    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2836        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2837        references. PCRE supports relative numbers with this syntax (I don't think
2838        Oniguruma does).
2839    
2840    12. Previously, a group with a zero repeat such as (...){0} was completely
2841        omitted from the compiled regex. However, this means that if the group
2842        was called as a subroutine from elsewhere in the pattern, things went wrong
2843        (an internal error was given). Such groups are now left in the compiled
2844        pattern, with a new opcode that causes them to be skipped at execution
2845        time.
2846    
2847    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2848        to the way PCRE behaves:
2849    
2850        (a) A lone ] character is dis-allowed (Perl treats it as data).
2851    
2852        (b) A back reference to an unmatched subpattern matches an empty string
2853            (Perl fails the current match path).
2854    
2855        (c) A data ] in a character class must be notated as \] because if the
2856            first data character in a class is ], it defines an empty class. (In
2857            Perl it is not possible to have an empty class.) The empty class []
2858            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2859            The negative empty class [^] matches any one character, independently
2860            of the DOTALL setting.
2861    
2862    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2863        non-existent subpattern following a character class starting with ']' and
2864        containing () gave an internal compiling error instead of "reference to
2865        non-existent subpattern". Fortunately, when the pattern did exist, the
2866        compiled code was correct. (When scanning forwards to check for the
2867        existence of the subpattern, it was treating the data ']' as terminating
2868        the class, so got the count wrong. When actually compiling, the reference
2869        was subsequently set up correctly.)
2870    
2871    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2872        it was being rejected as not supported by pcre_dfa_exec(), even though
2873        other assertions are supported. I have made pcre_dfa_exec() support
2874        (*FAIL).
2875    
2876    16. The implementation of 13c above involved the invention of a new opcode,
2877        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2878        cannot be changed at match time, I realized I could make a small
2879        improvement to matching performance by compiling OP_ALLANY instead of
2880        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2881        on the OP_ANY path.
2882    
2883    17. Compiling pcretest on Windows with readline support failed without the
2884        following two fixes: (1) Make the unistd.h include conditional on
2885        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2886    
2887    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2888        ncurses library to be included for pcretest when ReadLine support is
2889        requested, but also to allow for it to be overridden. This patch came from
2890        Daniel Bergström.
2891    
2892    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2893        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2894        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2895        spotting this.
2896    
2897    
2898  Version 7.6 28-Jan-08  Version 7.6 28-Jan-08
2899  ---------------------  ---------------------
2900    

Legend:
Removed from v.319  
changed lines
  Added in v.1691

  ViewVC Help
Powered by ViewVC 1.1.5