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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.341  
changed lines
  Added in v.1618

  ViewVC Help
Powered by ViewVC 1.1.5