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

Diff of /code/trunk/ChangeLog

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

revision 167 by ph10, Wed May 9 15:53:54 2007 UTC revision 1696 by ph10, Fri Apr 14 11:45:57 2017 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.2 01-May-07  Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5    development is happening in the PCRE2 10.xx series.
6    
7    Version 8.41
8    ------------
9    
10    1.  Fixed typo in CMakeLists.txt (wrong number of arguments for
11    PCRE_STATIC_RUNTIME (affects MSVC only).
12    
13    2.  Issue 1 for 8.40 below was not correctly fixed. If pcregrep in multiline
14    mode with --only-matching matched several lines, it restarted scanning at the
15    next line instead of moving on to the end of the matched string, which can be
16    several lines after the start.
17    
18    3.  Fix a missing else in the JIT compiler reported by 'idaifish'.
19    
20    4.  A (?# style comment is now ignored between a basic quantifier and a
21    following '+' or '?' (example: /X+(?#comment)?Y/.
22    
23    5.  Avoid use of a potentially overflowing buffer in pcregrep (patch by Petr
24    Pisar).
25    
26    6.  Fuzzers have reported issues in pcretest. These are NOT serious (it is,
27    after all, just a test program). However, to stop the reports, some easy ones
28    are fixed:
29    
30        (a) Check for values < 256 when calling isprint() in pcretest.
31        (b) Give an error for too big a number after \O.
32    
33    7.  In the 32-bit library in non-UTF mode, an attempt to find a Unicode
34    property for a character with a code point greater than 0x10ffff (the Unicode
35    maximum) caused a crash.
36    
37    8. The alternative matching function, pcre_dfa_exec() misbehaved if it
38    encountered a character class with a possessive repeat, for example [a-f]{3}+.
39    
40    9. When pcretest called pcre_copy_substring() in 32-bit mode, it set the buffer
41    length incorrectly, which could result in buffer overflow.
42    
43    10. Remove redundant line of code (accidentally left in ages ago).
44    
45    
46    Version 8.40 11-January-2017
47    ----------------------------
48    
49    1.  Using -o with -M in pcregrep could cause unnecessary repeated output when
50        the match extended over a line boundary.
51    
52    2.  Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
53        MSVC static compilation, putting the first patch under a new option.
54    
55    3.  Fix register overwite in JIT when SSE2 acceleration is enabled.
56    
57    4.  Ignore "show all captures" (/=) for DFA matching.
58    
59    5.  Fix JIT unaligned accesses on x86. Patch by Marc Mutz.
60    
61    6.  In any wide-character mode (8-bit UTF or any 16-bit or 32-bit mode),
62        without PCRE_UCP set, a negative character type such as \D in a positive
63        class should cause all characters greater than 255 to match, whatever else
64        is in the class. There was a bug that caused this not to happen if a
65        Unicode property item was added to such a class, for example [\D\P{Nd}] or
66        [\W\pL].
67    
68    7.  When pcretest was outputing information from a callout, the caret indicator
69        for the current position in the subject line was incorrect if it was after
70        an escape sequence for a character whose code point was greater than
71        \x{ff}.
72    
73    8.  A pattern such as (?<RA>abc)(?(R)xyz) was incorrectly compiled such that
74        the conditional was interpreted as a reference to capturing group 1 instead
75        of a test for recursion. Any group whose name began with R was
76        misinterpreted in this way. (The reference interpretation should only
77        happen if the group's name is precisely "R".)
78    
79    9.  A number of bugs have been mended relating to match start-up optimizations
80        when the first thing in a pattern is a positive lookahead. These all
81        applied only when PCRE_NO_START_OPTIMIZE was *not* set:
82    
83        (a) A pattern such as (?=.*X)X$ was incorrectly optimized as if it needed
84            both an initial 'X' and a following 'X'.
85        (b) Some patterns starting with an assertion that started with .* were
86            incorrectly optimized as having to match at the start of the subject or
87            after a newline. There are cases where this is not true, for example,
88            (?=.*[A-Z])(?=.{8,16})(?!.*[\s]) matches after the start in lines that
89            start with spaces. Starting .* in an assertion is no longer taken as an
90            indication of matching at the start (or after a newline).
91    
92    
93    Version 8.39 14-June-2016
94    -------------------------
95    
96    1.  If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
97        an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
98        misbehaved. This bug was found by the LLVM fuzzer.
99    
100    2.  Similar to the above, if an isolated \E was present between an item and its
101        qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
102        bug was found by the LLVM fuzzer.
103    
104    3.  Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
105        working correctly in UCP mode.
106    
107    4.  The POSIX wrapper function regexec() crashed if the option REG_STARTEND
108        was set when the pmatch argument was NULL. It now returns REG_INVARG.
109    
110    5.  Allow for up to 32-bit numbers in the ordin() function in pcregrep.
111    
112    6.  An empty \Q\E sequence between an item and its qualifier caused
113        pcre_compile() to misbehave when auto callouts were enabled. This bug was
114        found by the LLVM fuzzer.
115    
116    7.  If a pattern that was compiled with PCRE_EXTENDED started with white
117        space or a #-type comment that was followed by (?-x), which turns off
118        PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
119        pcre_compile() assumed that (?-x) applied to the whole pattern and
120        consequently mis-compiled it. This bug was found by the LLVM fuzzer.
121    
122    8.  A call of pcre_copy_named_substring() for a named substring whose number
123        was greater than the space in the ovector could cause a crash.
124    
125    9.  Yet another buffer overflow bug involved duplicate named groups with a
126        group that reset capture numbers (compare 8.38/7 below). Once again, I have
127        just allowed for more memory, even if not needed. (A proper fix is
128        implemented in PCRE2, but it involves a lot of refactoring.)
129    
130    10. pcre_get_substring_list() crashed if the use of \K in a match caused the
131        start of the match to be earlier than the end.
132    
133    11. Migrating appropriate PCRE2 JIT improvements to PCRE.
134    
135    12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
136        assertion, caused pcretest to generate incorrect output, and also to read
137        uninitialized memory (detected by ASAN or valgrind).
138    
139    13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
140        nested set of parentheses of sufficient size caused an overflow of the
141        compiling workspace (which was diagnosed, but of course is not desirable).
142    
143    14. And yet another buffer overflow bug involving duplicate named groups, this
144        time nested, with a nested back reference. Yet again, I have just allowed
145        for more memory, because anything more needs all the refactoring that has
146        been done for PCRE2. An example pattern that provoked this bug is:
147        /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
148        registered as CVE-2016-1283.
149    
150    15. pcretest went into a loop if global matching was requested with an ovector
151        size less than 2. It now gives an error message. This bug was found by
152        afl-fuzz.
153    
154    16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
155        ("assertion expected") when (?(?C) was not followed by an opening
156        parenthesis.
157    
158    17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
159        actually affect anything, by sheer luck.
160    
161    18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
162        static compilation.
163    
164    19. Modified the RunTest script to incorporate a valgrind suppressions file so
165        that certain errors, provoked by the SSE2 instruction set when JIT is used,
166        are ignored.
167    
168    20. A racing condition is fixed in JIT reported by Mozilla.
169    
170    21. Minor code refactor to avoid "array subscript is below array bounds"
171        compiler warning.
172    
173    22. Minor code refactor to avoid "left shift of negative number" warning.
174    
175    23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
176        UCP support.
177    
178    24. Refactor to avoid compiler warnings in pcrecpp.cc.
179    
180    25. Refactor to fix a typo in pcre_jit_test.c
181    
182    26. Patch to support compiling pcrecpp.cc with Intel compiler.
183    
184    
185    Version 8.38 23-November-2015
186    -----------------------------
187    
188    1.  If a group that contained a recursive back reference also contained a
189        forward reference subroutine call followed by a non-forward-reference
190        subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
191        compile correct code, leading to undefined behaviour or an internally
192        detected error. This bug was discovered by the LLVM fuzzer.
193    
194    2.  Quantification of certain items (e.g. atomic back references) could cause
195        incorrect code to be compiled when recursive forward references were
196        involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
197        This bug was discovered by the LLVM fuzzer.
198    
199    3.  A repeated conditional group whose condition was a reference by name caused
200        a buffer overflow if there was more than one group with the given name.
201        This bug was discovered by the LLVM fuzzer.
202    
203    4.  A recursive back reference by name within a group that had the same name as
204        another group caused a buffer overflow. For example:
205        /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
206    
207    5.  A forward reference by name to a group whose number is the same as the
208        current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
209        a buffer overflow at compile time. This bug was discovered by the LLVM
210        fuzzer.
211    
212    6.  A lookbehind assertion within a set of mutually recursive subpatterns could
213        provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
214    
215    7.  Another buffer overflow bug involved duplicate named groups with a
216        reference between their definition, with a group that reset capture
217        numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
218        fixed by always allowing for more memory, even if not needed. (A proper fix
219        is implemented in PCRE2, but it involves more refactoring.)
220    
221    8.  There was no check for integer overflow in subroutine calls such as (?123).
222    
223    9.  The table entry for \l in EBCDIC environments was incorrect, leading to its
224        being treated as a literal 'l' instead of causing an error.
225    
226    10. There was a buffer overflow if pcre_exec() was called with an ovector of
227        size 1. This bug was found by american fuzzy lop.
228    
229    11. If a non-capturing group containing a conditional group that could match
230        an empty string was repeated, it was not identified as matching an empty
231        string itself. For example: /^(?:(?(1)x|)+)+$()/.
232    
233    12. In an EBCDIC environment, pcretest was mishandling the escape sequences
234        \a and \e in test subject lines.
235    
236    13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
237        instead of the EBCDIC value.
238    
239    14. The handling of \c in an EBCDIC environment has been revised so that it is
240        now compatible with the specification in Perl's perlebcdic page.
241    
242    15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
243        ASCII/Unicode. This has now been added to the list of characters that are
244        recognized as white space in EBCDIC.
245    
246    16. When PCRE was compiled without UCP support, the use of \p and \P gave an
247        error (correctly) when used outside a class, but did not give an error
248        within a class.
249    
250    17. \h within a class was incorrectly compiled in EBCDIC environments.
251    
252    18. A pattern with an unmatched closing parenthesis that contained a backward
253        assertion which itself contained a forward reference caused buffer
254        overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
255    
256    19. JIT should return with error when the compiled pattern requires more stack
257        space than the maximum.
258    
259    20. A possessively repeated conditional group that could match an empty string,
260        for example, /(?(R))*+/, was incorrectly compiled.
261    
262    21. Fix infinite recursion in the JIT compiler when certain patterns such as
263        /(?:|a|){100}x/ are analysed.
264    
265    22. Some patterns with character classes involving [: and \\ were incorrectly
266        compiled and could cause reading from uninitialized memory or an incorrect
267        error diagnosis.
268    
269    23. Pathological patterns containing many nested occurrences of [: caused
270        pcre_compile() to run for a very long time.
271    
272    24. A conditional group with only one branch has an implicit empty alternative
273        branch and must therefore be treated as potentially matching an empty
274        string.
275    
276    25. If (?R was followed by - or + incorrect behaviour happened instead of a
277        diagnostic.
278    
279    26. Arrange to give up on finding the minimum matching length for overly
280        complex patterns.
281    
282    27. Similar to (4) above: in a pattern with duplicated named groups and an
283        occurrence of (?| it is possible for an apparently non-recursive back
284        reference to become recursive if a later named group with the relevant
285        number is encountered. This could lead to a buffer overflow. Wen Guanxing
286        from Venustech ADLAB discovered this bug.
287    
288    28. If pcregrep was given the -q option with -c or -l, or when handling a
289        binary file, it incorrectly wrote output to stdout.
290    
291    29. The JIT compiler did not restore the control verb head in case of *THEN
292        control verbs. This issue was found by Karl Skomski with a custom LLVM
293        fuzzer.
294    
295    30. Error messages for syntax errors following \g and \k were giving inaccurate
296        offsets in the pattern.
297    
298    31. Added a check for integer overflow in conditions (?(<digits>) and
299        (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
300        fuzzer.
301    
302    32. Handling recursive references such as (?2) when the reference is to a group
303        later in the pattern uses code that is very hacked about and error-prone.
304        It has been re-written for PCRE2. Here in PCRE1, a check has been added to
305        give an internal error if it is obvious that compiling has gone wrong.
306    
307    33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
308        This issue was found by Karl Skomski with a custom LLVM fuzzer.
309    
310    34. The JIT compiler should restore the control chain for empty possessive
311        repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
312    
313    35. Match limit check added to JIT recursion. This issue was found by Karl
314        Skomski with a custom LLVM fuzzer.
315    
316    36. Yet another case similar to 27 above has been circumvented by an
317        unconditional allocation of extra memory. This issue is fixed "properly" in
318        PCRE2 by refactoring the way references are handled. Wen Guanxing
319        from Venustech ADLAB discovered this bug.
320    
321    37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
322        with a custom LLVM fuzzer.
323    
324    38. Fixed a corner case of range optimization in JIT.
325    
326    39. An incorrect error "overran compiling workspace" was given if there were
327        exactly enough group forward references such that the last one extended
328        into the workspace safety margin. The next one would have expanded the
329        workspace. The test for overflow was not including the safety margin.
330    
331    40. A match limit issue is fixed in JIT which was found by Karl Skomski
332        with a custom LLVM fuzzer.
333    
334    41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
335        work under Windows. (Why has it taken so long for anyone to notice?)
336    
337    42. In a character class such as [\W\p{Any}] where both a negative-type escape
338        ("not a word character") and a property escape were present, the property
339        escape was being ignored.
340    
341    43. Fix crash caused by very long (*MARK) or (*THEN) names.
342    
343    44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
344        by a single ASCII character in a class item, was incorrectly compiled in
345        UCP mode. The POSIX class got lost, but only if the single character
346        followed it.
347    
348    45. [:punct:] in UCP mode was matching some characters in the range 128-255
349        that should not have been matched.
350    
351    46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
352        class, all characters with code points greater than 255 are in the class.
353        When a Unicode property was also in the class (if PCRE_UCP is set, escapes
354        such as \w are turned into Unicode properties), wide characters were not
355        correctly handled, and could fail to match.
356    
357    
358    Version 8.37 28-April-2015
359    --------------------------
360    
361    1.  When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
362        for those parentheses to be closed with whatever has been captured so far.
363        However, it was failing to mark any other groups between the hightest
364        capture so far and the currrent group as "unset". Thus, the ovector for
365        those groups contained whatever was previously there. An example is the
366        pattern /(x)|((*ACCEPT))/ when matched against "abcd".
367    
368    2.  If an assertion condition was quantified with a minimum of zero (an odd
369        thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
370    
371    3.  If a pattern in pcretest input had the P (POSIX) modifier followed by an
372        unrecognized modifier, a crash could occur.
373    
374    4.  An attempt to do global matching in pcretest with a zero-length ovector
375        caused a crash.
376    
377    5.  Fixed a memory leak during matching that could occur for a subpattern
378        subroutine call (recursive or otherwise) if the number of captured groups
379        that had to be saved was greater than ten.
380    
381    6.  Catch a bad opcode during auto-possessification after compiling a bad UTF
382        string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
383        UTF with NO_UTF_CHECK is documented as having an undefined outcome.
384    
385    7.  A UTF pattern containing a "not" match of a non-ASCII character and a
386        subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
387    
388    8. When a pattern is compiled, it remembers the highest back reference so that
389       when matching, if the ovector is too small, extra memory can be obtained to
390       use instead. A conditional subpattern whose condition is a check on a
391       capture having happened, such as, for example in the pattern
392       /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
393       setting the highest backreference number. This mattered only if pcre_exec()
394       was called with an ovector that was too small to hold the capture, and there
395       was no other kind of back reference (a situation which is probably quite
396       rare). The effect of the bug was that the condition was always treated as
397       FALSE when the capture could not be consulted, leading to a incorrect
398       behaviour by pcre_exec(). This bug has been fixed.
399    
400    9. A reference to a duplicated named group (either a back reference or a test
401       for being set in a conditional) that occurred in a part of the pattern where
402       PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
403       to be incorrectly calculated, leading to overwriting.
404    
405    10. A mutually recursive set of back references such as (\2)(\1) caused a
406        segfault at study time (while trying to find the minimum matching length).
407        The infinite loop is now broken (with the minimum length unset, that is,
408        zero).
409    
410    11. If an assertion that was used as a condition was quantified with a minimum
411        of zero, matching went wrong. In particular, if the whole group had
412        unlimited repetition and could match an empty string, a segfault was
413        likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
414        assertions to be quantified, but not if they are being used as conditions,
415        so the above pattern is faulted by Perl. PCRE has now been changed so that
416        it also rejects such patterns.
417    
418    12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
419        failed to allow the zero-repeat case if pcre2_exec() was called with an
420        ovector too small to capture the group.
421    
422    13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
423        Red Hat Product Security:
424    
425        (a) A crash if /K and /F were both set with the option to save the compiled
426        pattern.
427    
428        (b) Another crash if the option to print captured substrings in a callout
429        was combined with setting a null ovector, for example \O\C+ as a subject
430        string.
431    
432    14. A pattern such as "((?2){0,1999}())?", which has a group containing a
433        forward reference repeated a large (but limited) number of times within a
434        repeated outer group that has a zero minimum quantifier, caused incorrect
435        code to be compiled, leading to the error "internal error:
436        previously-checked referenced subpattern not found" when an incorrect
437        memory address was read. This bug was reported as "heap overflow",
438        discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
439        CVE-2015-2325.
440    
441    23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
442        call within a group that also contained a recursive back reference caused
443        incorrect code to be compiled. This bug was reported as "heap overflow",
444        discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
445        number CVE-2015-2326.
446    
447    24. Computing the size of the JIT read-only data in advance has been a source
448        of various issues, and new ones are still appear unfortunately. To fix
449        existing and future issues, size computation is eliminated from the code,
450        and replaced by on-demand memory allocation.
451    
452    25. A pattern such as /(?i)[A-`]/, where characters in the other case are
453        adjacent to the end of the range, and the range contained characters with
454        more than one other case, caused incorrect behaviour when compiled in UTF
455        mode. In that example, the range a-j was left out of the class.
456    
457    26. Fix JIT compilation of conditional blocks, which assertion
458        is converted to (*FAIL). E.g: /(?(?!))/.
459    
460    27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
461        discovered by the LLVM fuzzer.
462    
463    28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
464        when this assertion was used as a condition, for example (?(?!)a|b). In
465        pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
466        error about an unsupported item.
467    
468    29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
469        possessification code could take exponential time to complete. A recursion
470        depth limit of 1000 has been imposed to limit the resources used by this
471        optimization.
472    
473    30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
474        such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
475        because \S ensures they are all in the class. The code for doing this was
476        interacting badly with the code for computing the amount of space needed to
477        compile the pattern, leading to a buffer overflow. This bug was discovered
478        by the LLVM fuzzer.
479    
480    31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
481        other kinds of group caused stack overflow at compile time. This bug was
482        discovered by the LLVM fuzzer.
483    
484    32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
485        between a subroutine call and its quantifier was incorrectly compiled,
486        leading to buffer overflow or other errors. This bug was discovered by the
487        LLVM fuzzer.
488    
489    33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
490        assertion after (?(. The code was failing to check the character after
491        (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
492        was discovered by the LLVM fuzzer.
493    
494    34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
495        a fixed maximum following a group that contains a subroutine reference was
496        incorrectly compiled and could trigger buffer overflow. This bug was
497        discovered by the LLVM fuzzer.
498    
499    35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
500        caused a stack overflow instead of the diagnosis of a non-fixed length
501        lookbehind assertion. This bug was discovered by the LLVM fuzzer.
502    
503    36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
504        (e.g. /(?<=\Ka)/) could make pcregrep loop.
505    
506    37. There was a similar problem to 36 in pcretest for global matches.
507    
508    38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
509        and a subsequent item in the pattern caused a non-match, backtracking over
510        the repeated \X did not stop, but carried on past the start of the subject,
511        causing reference to random memory and/or a segfault. There were also some
512        other cases where backtracking after \C could crash. This set of bugs was
513        discovered by the LLVM fuzzer.
514    
515    39. The function for finding the minimum length of a matching string could take
516        a very long time if mutual recursion was present many times in a pattern,
517        for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
518        method has been implemented. This infelicity was discovered by the LLVM
519        fuzzer.
520    
521    40. Static linking against the PCRE library using the pkg-config module was
522        failing on missing pthread symbols.
523    
524    
525    Version 8.36 26-September-2014
526    ------------------------------
527    
528    1.  Got rid of some compiler warnings in the C++ modules that were shown up by
529        -Wmissing-field-initializers and -Wunused-parameter.
530    
531    2.  The tests for quantifiers being too big (greater than 65535) were being
532        applied after reading the number, and stupidly assuming that integer
533        overflow would give a negative number. The tests are now applied as the
534        numbers are read.
535    
536    3.  Tidy code in pcre_exec.c where two branches that used to be different are
537        now the same.
538    
539    4.  The JIT compiler did not generate match limit checks for certain
540        bracketed expressions with quantifiers. This may lead to exponential
541        backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
542        issue should be resolved now.
543    
544    5.  Fixed an issue, which occures when nested alternatives are optimized
545        with table jumps.
546    
547    6.  Inserted two casts and changed some ints to size_t in the light of some
548        reported 64-bit compiler warnings (Bugzilla 1477).
549    
550    7.  Fixed a bug concerned with zero-minimum possessive groups that could match
551        an empty string, which sometimes were behaving incorrectly in the
552        interpreter (though correctly in the JIT matcher). This pcretest input is
553        an example:
554    
555          '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
556          NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
557    
558        the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
559        JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
560        for an empty string was breaking the inner loop and carrying on at a lower
561        level, when possessive repeated groups should always return to a higher
562        level as they have no backtrack points in them. The empty string test now
563        occurs at the outer level.
564    
565    8.  Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
566        ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
567    
568    9.  Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
569        doesn't).
570    
571    10. Change 8.34/15 introduced a bug that caused the amount of memory needed
572        to hold a pattern to be incorrectly computed (too small) when there were
573        named back references to duplicated names. This could cause "internal
574        error: code overflow" or "double free or corruption" or other memory
575        handling errors.
576    
577    11. When named subpatterns had the same prefixes, back references could be
578        confused. For example, in this pattern:
579    
580          /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
581    
582        the reference to 'Name' was incorrectly treated as a reference to a
583        duplicate name.
584    
585    12. A pattern such as /^s?c/mi8 where the optional character has more than
586        one "other case" was incorrectly compiled such that it would only try to
587        match starting at "c".
588    
589    13. When a pattern starting with \s was studied, VT was not included in the
590        list of possible starting characters; this should have been part of the
591        8.34/18 patch.
592    
593    14. If a character class started [\Qx]... where x is any character, the class
594        was incorrectly terminated at the ].
595    
596    15. If a pattern that started with a caseless match for a character with more
597        than one "other case" was studied, PCRE did not set up the starting code
598        unit bit map for the list of possible characters. Now it does. This is an
599        optimization improvement, not a bug fix.
600    
601    16. The Unicode data tables have been updated to Unicode 7.0.0.
602    
603    17. Fixed a number of memory leaks in pcregrep.
604    
605    18. Avoid a compiler warning (from some compilers) for a function call with
606        a cast that removes "const" from an lvalue by using an intermediate
607        variable (to which the compiler does not object).
608    
609    19. Incorrect code was compiled if a group that contained an internal recursive
610        back reference was optional (had quantifier with a minimum of zero). This
611        example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
612        caused segmentation faults because of stack overflows at compile time.
613    
614    20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
615        group that is quantified with an indefinite repeat, caused a compile-time
616        loop which used up all the system stack and provoked a segmentation fault.
617        This was not the same bug as 19 above.
618    
619    21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
620        Patch by Mike Frysinger.
621    
622    
623    Version 8.35 04-April-2014
624    --------------------------
625    
626    1.  A new flag is set, when property checks are present in an XCLASS.
627        When this flag is not set, PCRE can perform certain optimizations
628        such as studying these XCLASS-es.
629    
630    2.  The auto-possessification of character sets were improved: a normal
631        and an extended character set can be compared now. Furthermore
632        the JIT compiler optimizes more character set checks.
633    
634    3.  Got rid of some compiler warnings for potentially uninitialized variables
635        that show up only when compiled with -O2.
636    
637    4.  A pattern such as (?=ab\K) that uses \K in an assertion can set the start
638        of a match later then the end of the match. The pcretest program was not
639        handling the case sensibly - it was outputting from the start to the next
640        binary zero. It now reports this situation in a message, and outputs the
641        text from the end to the start.
642    
643    5.  Fast forward search is improved in JIT. Instead of the first three
644        characters, any three characters with fixed position can be searched.
645        Search order: first, last, middle.
646    
647    6.  Improve character range checks in JIT. Characters are read by an inprecise
648        function now, which returns with an unknown value if the character code is
649        above a certain threshold (e.g: 256). The only limitation is that the value
650        must be bigger than the threshold as well. This function is useful when
651        the characters above the threshold are handled in the same way.
652    
653    7.  The macros whose names start with RAWUCHAR are placeholders for a future
654        mode in which only the bottom 21 bits of 32-bit data items are used. To
655        make this more memorable for those maintaining the code, the names have
656        been changed to start with UCHAR21, and an extensive comment has been added
657        to their definition.
658    
659    8.  Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
660        to the export list in Makefile.am (they were accidentally omitted from the
661        8.34 tarball).
662    
663    9.  The informational output from pcretest used the phrase "starting byte set"
664        which is inappropriate for the 16-bit and 32-bit libraries. As the output
665        for "first char" and "need char" really means "non-UTF-char", I've changed
666        "byte" to "char", and slightly reworded the output. The documentation about
667        these values has also been (I hope) clarified.
668    
669    10. Another JIT related optimization: use table jumps for selecting the correct
670        backtracking path, when more than four alternatives are present inside a
671        bracket.
672    
673    11. Empty match is not possible, when the minimum length is greater than zero,
674        and there is no \K in the pattern. JIT should avoid empty match checks in
675        such cases.
676    
677    12. In a caseless character class with UCP support, when a character with more
678        than one alternative case was not the first character of a range, not all
679        the alternative cases were added to the class. For example, s and \x{17f}
680        are both alternative cases for S: the class [RST] was handled correctly,
681        but [R-T] was not.
682    
683    13. The configure.ac file always checked for pthread support when JIT was
684        enabled. This is not used in Windows, so I have put this test inside a
685        check for the presence of windows.h (which was already tested for).
686    
687    14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
688        The algorithm provides a way to skip certain starting offsets, and usually
689        faster than linear prefix searches.
690    
691    15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
692        as for 'fr_FR' and 'french'. For some reason, however, it then used the
693        Windows-specific input and output files, which have 'french' screwed in.
694        So this could never have worked. One of the problems with locales is that
695        they aren't always the same. I have now updated RunTest so that it checks
696        the output of the locale test (test 3) against three different output
697        files, and it allows the test to pass if any one of them matches. With luck
698        this should make the test pass on some versions of Solaris where it was
699        failing. Because of the uncertainty, the script did not used to stop if
700        test 3 failed; it now does. If further versions of a French locale ever
701        come to light, they can now easily be added.
702    
703    16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
704        there was a message during ./configure, but it did not stop. This now
705        provokes an error. The invalid example in README has been corrected.
706        If a value less than the minimum is given, the minimum value has always
707        been used, but now a warning is given.
708    
709    17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
710        was a bug in the test system, which is now fixed. Also, the list of various
711        configurations that are tested for each release did not have one with both
712        16/32 bits and --enable-bar-anycrlf. It now does.
713    
714    18. pcretest was missing "-C bsr" for displaying the \R default setting.
715    
716    19. Little endian PowerPC systems are supported now by the JIT compiler.
717    
718    20. The fast forward newline mechanism could enter to an infinite loop on
719        certain invalid UTF-8 input. Although we don't support these cases
720        this issue can be fixed by a performance optimization.
721    
722    21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
723        not take account if existing stack usage. There is now a new global
724        variable called pcre_stack_guard that can be set to point to an external
725        function to check stack availability. It is called at the start of
726        processing every parenthesized group.
727    
728    22. A typo in the code meant that in ungreedy mode the max/min qualifier
729        behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
730        not match "ab".
731    
732    23. When UTF was disabled, the JIT program reported some incorrect compile
733        errors. These messages are silenced now.
734    
735    24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
736        compiler.
737    
738    25. Change all the temporary files used in RunGrepTest to be different to those
739        used by RunTest so that the tests can be run simultaneously, for example by
740        "make -j check".
741    
742    
743    Version 8.34 15-December-2013
744    -----------------------------
745    
746    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
747        executable memory. Patch inspired by Carsten Klein.
748    
749    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
750        this macro is never tested and has no effect, because the work to support
751        coverage involves only compiling and linking options and special targets in
752        the Makefile. The comment in config.h implied that defining the macro would
753        enable coverage support, which is totally false. There was also support for
754        setting this macro in the CMake files (my fault, I just copied it from
755        configure). SUPPORT_GCOV has now been removed.
756    
757    3.  Make a small performance improvement in strlen16() and strlen32() in
758        pcretest.
759    
760    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
761        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
762        these cases). There was also one in pcretest.c.
763    
764    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
765    
766    6.  In UTF mode, the code for checking whether a group could match an empty
767        string (which is used for indefinitely repeated groups to allow for
768        breaking an infinite loop) was broken when the group contained a repeated
769        negated single-character class with a character that occupied more than one
770        data item and had a minimum repetition of zero (for example, [^\x{100}]* in
771        UTF-8 mode). The effect was undefined: the group might or might not be
772        deemed as matching an empty string, or the program might have crashed.
773    
774    7.  The code for checking whether a group could match an empty string was not
775        recognizing that \h, \H, \v, \V, and \R must match a character.
776    
777    8.  Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
778        an empty string. If it can, pcretest shows this in its information output.
779    
780    9.  Fixed two related bugs that applied to Unicode extended grapheme clusters
781        that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
782        matched by pcre_exec() without using JIT:
783    
784        (a) If the rest of the pattern did not match after a maximal run of
785            grapheme clusters, the code for backing up to try with fewer of them
786            did not always back up over a full grapheme when characters that do not
787            have the modifier quality were involved, e.g. Hangul syllables.
788    
789        (b) If the match point in a subject started with modifier character, and
790            there was no match, the code could incorrectly back up beyond the match
791            point, and potentially beyond the first character in the subject,
792            leading to a segfault or an incorrect match result.
793    
794    10. A conditional group with an assertion condition could lead to PCRE
795        recording an incorrect first data item for a match if no other first data
796        item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
797        first data item, and therefore matched "ca" after "c" instead of at the
798        start.
799    
800    11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
801        bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
802    
803    12. The source of pcregrep now includes z/OS-specific code so that it can be
804        compiled for z/OS as part of the special z/OS distribution.
805    
806    13. Added the -T and -TM options to pcretest.
807    
808    14. The code in pcre_compile.c for creating the table of named capturing groups
809        has been refactored. Instead of creating the table dynamically during the
810        actual compiling pass, the information is remembered during the pre-compile
811        pass (on the stack unless there are more than 20 named groups, in which
812        case malloc() is used) and the whole table is created before the actual
813        compile happens. This has simplified the code (it is now nearly 150 lines
814        shorter) and prepared the way for better handling of references to groups
815        with duplicate names.
816    
817    15. A back reference to a named subpattern when there is more than one of the
818        same name now checks them in the order in which they appear in the pattern.
819        The first one that is set is used for the reference. Previously only the
820        first one was inspected. This change makes PCRE more compatible with Perl.
821    
822    16. Unicode character properties were updated from Unicode 6.3.0.
823    
824    17. The compile-time code for auto-possessification has been refactored, based
825        on a patch by Zoltan Herczeg. It now happens after instead of during
826        compilation. The code is cleaner, and more cases are handled. The option
827        PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
828        options in pcretest are provided to set it. It can also be set by
829        (*NO_AUTO_POSSESS) at the start of a pattern.
830    
831    18. The character VT has been added to the default ("C" locale) set of
832        characters that match \s and are generally treated as white space,
833        following this same change in Perl 5.18. There is now no difference between
834        "Perl space" and "POSIX space". Whether VT is treated as white space in
835        other locales depends on the locale.
836    
837    19. The code for checking named groups as conditions, either for being set or
838        for being recursed, has been refactored (this is related to 14 and 15
839        above). Processing unduplicated named groups should now be as fast at
840        numerical groups, and processing duplicated groups should be faster than
841        before.
842    
843    20. Two patches to the CMake build system, by Alexander Barkov:
844    
845          (1) Replace the "source" command by "." in CMakeLists.txt because
846              "source" is a bash-ism.
847    
848          (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
849              without these the CMake build does not work on Solaris.
850    
851    21. Perl has changed its handling of \8 and \9. If there is no previously
852        encountered capturing group of those numbers, they are treated as the
853        literal characters 8 and 9 instead of a binary zero followed by the
854        literals. PCRE now does the same.
855    
856    22. Following Perl, added \o{} to specify codepoints in octal, making it
857        possible to specify values greater than 0777 and also making them
858        unambiguous.
859    
860    23. Perl now gives an error for missing closing braces after \x{... instead of
861        treating the string as literal. PCRE now does the same.
862    
863    24. RunTest used to grumble if an inappropriate test was selected explicitly,
864        but just skip it when running all tests. This make it awkward to run ranges
865        of tests when one of them was inappropriate. Now it just skips any
866        inappropriate tests, as it always did when running all tests.
867    
868    25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
869        character types such as \d or \w, too many callouts were inserted, and the
870        data that they returned was rubbish.
871    
872    26. In UCP mode, \s was not matching two of the characters that Perl matches,
873        namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
874        were matched by \h. The code has now been refactored so that the lists of
875        the horizontal and vertical whitespace characters used for \h and \v (which
876        are defined only in one place) are now also used for \s.
877    
878    27. Add JIT support for the 64 bit TileGX architecture.
879        Patch by Jiong Wang (Tilera Corporation).
880    
881    28. Possessive quantifiers for classes (both explicit and automatically
882        generated) now use special opcodes instead of wrapping in ONCE brackets.
883    
884    29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
885        (because it's meaningless), this was not happening when PCRE_CASELESS was
886        set. Not wrong, but inefficient.
887    
888    30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
889        properties for \w, \d, etc) is present in a test regex. Otherwise if the
890        test contains no characters greater than 255, Perl doesn't realise it
891        should be using Unicode semantics.
892    
893    31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
894        [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
895        does in Unicode mode.
896    
897    32. Added the "forbid" facility to pcretest so that putting tests into the
898        wrong test files can sometimes be quickly detected.
899    
900    33. There is now a limit (default 250) on the depth of nesting of parentheses.
901        This limit is imposed to control the amount of system stack used at compile
902        time. It can be changed at build time by --with-parens-nest-limit=xxx or
903        the equivalent in CMake.
904    
905    34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
906        errors. Perl warns for these when in warning mode, but PCRE has no facility
907        for giving warnings.
908    
909    35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
910        However, this was not working for (?!) because it is optimized to (*FAIL),
911        for which PCRE does not allow quantifiers. The optimization is now disabled
912        when a quantifier follows (?!). I can't see any use for this, but it makes
913        things uniform.
914    
915    36. Perl no longer allows group names to start with digits, so I have made this
916        change also in PCRE. It simplifies the code a bit.
917    
918    37. In extended mode, Perl ignores spaces before a + that indicates a
919        possessive quantifier. PCRE allowed a space before the quantifier, but not
920        before the possessive +. It now does.
921    
922    38. The use of \K (reset reported match start) within a repeated possessive
923        group such as (a\Kb)*+ was not working.
924    
925    40. Document that the same character tables must be used at compile time and
926        run time, and that the facility to pass tables to pcre_exec() and
927        pcre_dfa_exec() is for use only with saved/restored patterns.
928    
929    41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
930        features for Builds with MSVC:
931    
932        1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
933           builds). The .rc files can be used to set FileDescription and many other
934           attributes.
935    
936        2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
937           This allows higher-level build scripts which want .pdb files to avoid
938           hard-coding the exact files needed."
939    
940    42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
941        mean "start of word" and "end of word", respectively, as a transition aid.
942    
943    43. A minimizing repeat of a class containing codepoints greater than 255 in
944        non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
945        compiled to use the heap for recursion.
946    
947    44. Got rid of some compiler warnings for unused variables when UTF but not UCP
948        is configured.
949    
950    
951    Version 8.33 28-May-2013
952    ------------------------
953    
954    1.  Added 'U' to some constants that are compared to unsigned integers, to
955        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
956        variables that are added to signed variables, to ensure the result is
957        signed and can be negated.
958    
959    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
960        CMake config files.
961    
962    3.  Revise the creation of config.h.generic so that all boolean macros are
963        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
964        overriding via -D on the command line possible.
965    
966    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
967        to unsigned int is reported to make a quite noticeable speed difference in
968        a specific Windows environment. Testing on Linux did also appear to show
969        some benefit (and it is clearly not harmful). Also fixed the definition of
970        Xop which should be unsigned.
971    
972    5.  Related to (4), changing the definition of the intermediate variable cc
973        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
974        improvements.
975    
976    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
977        unnecessary spaces.
978    
979    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
980        and later.
981    
982    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
983    
984    9.  Optimizing fast_forward_start_bits in JIT.
985    
986    10. Adding support for callouts in JIT, and fixing some issues revealed
987        during this work. Namely:
988    
989        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
990    
991        (b) Minimum length was not checked before the matching is started.
992    
993    11. The value of capture_last that is passed to callouts was incorrect in some
994        cases when there was a capture on one path that was subsequently abandoned
995        after a backtrack. Also, the capture_last value is now reset after a
996        recursion, since all captures are also reset in this case.
997    
998    12. The interpreter no longer returns the "too many substrings" error in the
999        case when an overflowing capture is in a branch that is subsequently
1000        abandoned after a backtrack.
1001    
1002    13. In the pathological case when an offset vector of size 2 is used, pcretest
1003        now prints out the matched string after a yield of 0 or 1.
1004    
1005    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
1006        Only supported by the JIT compiler at the moment.
1007    
1008    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
1009    
1010    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
1011        offset of the starting point of the matching process, provided the offsets
1012        vector is large enough.
1013    
1014    17. The \A escape now records a lookbehind value of 1, though its execution
1015        does not actually inspect the previous character. This is to ensure that,
1016        in partial multi-segment matching, at least one character from the old
1017        segment is retained when a new segment is processed. Otherwise, if there
1018        are no lookbehinds in the pattern, \A might match incorrectly at the start
1019        of a new segment.
1020    
1021    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
1022    
1023    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
1024        gives some modest performance improvement in 8-bit mode.
1025    
1026    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
1027        be expressed in certain programming languages using Universal Character
1028        Names.
1029    
1030    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
1031        which points out that "non characters" are not "characters that may not
1032        appear in Unicode strings" but rather "characters that are reserved for
1033        internal use and have only local meaning".
1034    
1035    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
1036        there was a conditional group that depended on an assertion, if the
1037        assertion was false, the callout that immediately followed the alternation
1038        in the condition was skipped when pcre_exec() was used for matching.
1039    
1040    23. Allow an explicit callout to be inserted before an assertion that is the
1041        condition for a conditional group, for compatibility with automatic
1042        callouts, which always insert a callout at this point.
1043    
1044    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
1045        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
1046    
1047    25. (*PRUNE) is now supported by the JIT compiler.
1048    
1049    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
1050    
1051    27. Fix the case where there are two or more SKIPs with arguments that may be
1052        ignored.
1053    
1054    28. (*SKIP) is now supported by the JIT compiler.
1055    
1056    29. (*THEN) is now supported by the JIT compiler.
1057    
1058    30. Update RunTest with additional test selector options.
1059    
1060    31. The way PCRE handles backtracking verbs has been changed in two ways.
1061    
1062        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
1063        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
1064        backtracking. In some cases this makes it more Perl-compatible, but Perl's
1065        rather obscure rules do not always do the same thing.
1066    
1067        (2) Previously, backtracking verbs were confined within assertions. This is
1068        no longer the case for positive assertions, except for (*ACCEPT). Again,
1069        this sometimes improves Perl compatibility, and sometimes does not.
1070    
1071    32. A number of tests that were in test 2 because Perl did things differently
1072        have been moved to test 1, because either Perl or PCRE has changed, and
1073        these tests are now compatible.
1074    
1075    32. Backtracking control verbs are now handled in the same way in JIT and
1076        interpreter.
1077    
1078    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1079        contained a forward subroutine reference caused a compile error.
1080    
1081    34. Auto-detect and optimize limited repetitions in JIT.
1082    
1083    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1084        blocking (*UTF) etc.
1085    
1086    36. In the interpreter, maximizing pattern repetitions for characters and
1087        character types now use tail recursion, which reduces stack usage.
1088    
1089    37. The value of the max lookbehind was not correctly preserved if a compiled
1090        and saved regex was reloaded on a host of different endianness.
1091    
1092    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1093        of the compiled pattern block, expand the flags field from 16 to 32 bits
1094        because it was almost full.
1095    
1096    39. Try madvise first before posix_madvise.
1097    
1098    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1099        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1100        original change locked out all matches of empty strings. This has been
1101        changed so that one match of an empty string per line is recognized.
1102        Subsequent searches on the same line (for colouring or for --only-matching,
1103        for example) do not recognize empty strings.
1104    
1105    41. Applied a user patch to fix a number of spelling mistakes in comments.
1106    
1107    42. Data lines longer than 65536 caused pcretest to crash.
1108    
1109    43. Clarified the data type for length and startoffset arguments for pcre_exec
1110        and pcre_dfa_exec in the function-specific man pages, where they were
1111        explicitly stated to be in bytes, never having been updated. I also added
1112        some clarification to the pcreapi man page.
1113    
1114    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1115        a segmentation fault.
1116    
1117    
1118    Version 8.32 30-November-2012
1119    -----------------------------
1120    
1121    1.  Improved JIT compiler optimizations for first character search and single
1122        character iterators.
1123    
1124    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1125        Patch by Daniel Richard G.
1126    
1127    3.  Single character iterator optimizations in the JIT compiler.
1128    
1129    4.  Improved JIT compiler optimizations for character ranges.
1130    
1131    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1132        Reported by Giuseppe D'Angelo.
1133    
1134    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1135        of a line, was being set incorrectly in cases where .* appeared inside
1136        atomic brackets at the start of a pattern, or where there was a subsequent
1137        *PRUNE or *SKIP.
1138    
1139    7.  Improved instruction cache flush for POWER/PowerPC.
1140        Patch by Daniel Richard G.
1141    
1142    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
1143        grep:
1144    
1145        (a) There is now no limit to the number of patterns to be matched.
1146    
1147        (b) An error is given if a pattern is too long.
1148    
1149        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1150            are now supported.
1151    
1152        (d) --exclude-from and --include-from (multiple use) have been added.
1153    
1154        (e) Exclusions and inclusions now apply to all files and directories, not
1155            just to those obtained from scanning a directory recursively.
1156    
1157        (f) Multiple uses of -f and --file-list are now supported.
1158    
1159        (g) In a Windows environment, the default for -d has been changed from
1160            "read" (the GNU grep default) to "skip", because otherwise the presence
1161            of a directory in the file list provokes an error.
1162    
1163        (h) The documentation has been revised and clarified in places.
1164    
1165    9.  Improve the matching speed of capturing brackets.
1166    
1167    10. Changed the meaning of \X so that it now matches a Unicode extended
1168        grapheme cluster.
1169    
1170    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1171        out POSIX threads when JIT support is configured.
1172    
1173    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1174    
1175    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1176        structure could go wrong in environments where size_t is not the same size
1177        as int.
1178    
1179    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1180        set.
1181    
1182    15. The EBCDIC support had decayed; later updates to the code had included
1183        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1184        general tidy up of EBCDIC-related issues, and the documentation was also
1185        not quite right. There is now a test that can be run on ASCII systems to
1186        check some of the EBCDIC-related things (but is it not a full test).
1187    
1188    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1189        in a small tidy to the code.
1190    
1191    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1192    
1193    18. If the --only-matching (-o) option in pcregrep is specified multiple
1194        times, each one causes appropriate output. For example, -o1 -o2 outputs the
1195        substrings matched by the 1st and 2nd capturing parentheses. A separating
1196        string can be specified by --om-separator (default empty).
1197    
1198    19. Improving the first n character searches.
1199    
1200    20. Turn case lists for horizontal and vertical white space into macros so that
1201        they are defined only once.
1202    
1203    21. This set of changes together give more compatible Unicode case-folding
1204        behaviour for characters that have more than one other case when UCP
1205        support is available.
1206    
1207        (a) The Unicode property table now has offsets into a new table of sets of
1208            three or more characters that are case-equivalent. The MultiStage2.py
1209            script that generates these tables (the pcre_ucd.c file) now scans
1210            CaseFolding.txt instead of UnicodeData.txt for character case
1211            information.
1212    
1213        (b) The code for adding characters or ranges of characters to a character
1214            class has been abstracted into a generalized function that also handles
1215            case-independence. In UTF-mode with UCP support, this uses the new data
1216            to handle characters with more than one other case.
1217    
1218        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1219            whose other case is greater than 256 are now correctly matched
1220            caselessly. Previously, the high codepoint matched the low one, but not
1221            vice versa.
1222    
1223        (d) The processing of \h, \H, \v, and \ in character classes now makes use
1224            of the new class addition function, using character lists defined as
1225            macros alongside the case definitions of 20 above.
1226    
1227        (e) Caseless back references now work with characters that have more than
1228            one other case.
1229    
1230        (f) General caseless matching of characters with more than one other case
1231            is supported.
1232    
1233    22. Unicode character properties were updated from Unicode 6.2.0
1234    
1235    23. Improved CMake support under Windows. Patch by Daniel Richard G.
1236    
1237    24. Add support for 32-bit character strings, and UTF-32
1238    
1239    25. Major JIT compiler update (code refactoring and bugfixing).
1240        Experimental Sparc 32 support is added.
1241    
1242    26. Applied a modified version of Daniel Richard G's patch to create
1243        pcre.h.generic and config.h.generic by "make" instead of in the
1244        PrepareRelease script.
1245    
1246    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1247        pcre_compile.c when checking for a zero character.
1248    
1249    28. Introducing a native interface for JIT. Through this interface, the compiled
1250        machine code can be directly executed. The purpose of this interface is to
1251        provide fast pattern matching, so several sanity checks are not performed.
1252        However, feature tests are still performed. The new interface provides
1253        1.4x speedup compared to the old one.
1254    
1255    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1256        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1257        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1258    
1259    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1260        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1261        the "old" RFC 2279). Instead, it ended up passing a negative length to
1262        pcre_exec().
1263    
1264    31. Add support for GCC's visibility feature to hide internal functions.
1265    
1266    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1267        "unknown -C option" after outputting 0 or 1.
1268    
1269    33. There is now support for generating a code coverage report for the test
1270        suite in environments where gcc is the compiler and lcov is installed. This
1271        is mainly for the benefit of the developers.
1272    
1273    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1274        unaddressable using valgrind annotations, allowing valgrind to detect
1275        invalid memory accesses. This is mainly for the benefit of the developers.
1276    
1277    25. (*UTF) can now be used to start a pattern in any of the three libraries.
1278    
1279    26. Give configure error if --enable-cpp but no C++ compiler found.
1280    
1281    
1282    Version 8.31 06-July-2012
1283    -------------------------
1284    
1285    1.  Fixing a wrong JIT test case and some compiler warnings.
1286    
1287    2.  Removed a bashism from the RunTest script.
1288    
1289    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1290        to unsigned type, result still unsigned" that was given by an MS compiler
1291        on encountering the code "-sizeof(xxx)".
1292    
1293    4.  Partial matching support is added to the JIT compiler.
1294    
1295    5.  Fixed several bugs concerned with partial matching of items that consist
1296        of more than one character:
1297    
1298        (a) /^(..)\1/ did not partially match "aba" because checking references was
1299            done on an "all or nothing" basis. This also applied to repeated
1300            references.
1301    
1302        (b) \R did not give a hard partial match if \r was found at the end of the
1303            subject.
1304    
1305        (c) \X did not give a hard partial match after matching one or more
1306            characters at the end of the subject.
1307    
1308        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1309            a partial match for the string "\r".
1310    
1311        (e) When newline was set to CRLF, the metacharacter "." did not recognize
1312            a partial match for a CR character at the end of the subject string.
1313    
1314    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1315        running pcretest, the text "(JIT)" added to the output whenever JIT is
1316        actually used to run the match.
1317    
1318    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
1319        or /S+[+] with a digit between 1 and 7.
1320    
1321    8.  OP_NOT now supports any UTF character not just single-byte ones.
1322    
1323    9.  (*MARK) control verb is now supported by the JIT compiler.
1324    
1325    10. The command "./RunTest list" lists the available tests without actually
1326        running any of them. (Because I keep forgetting what they all are.)
1327    
1328    11. Add PCRE_INFO_MAXLOOKBEHIND.
1329    
1330    12. Applied a (slightly modified) user-supplied patch that improves performance
1331        when the heap is used for recursion (compiled with --disable-stack-for-
1332        recursion). Instead of malloc and free for each heap frame each time a
1333        logical recursion happens, frames are retained on a chain and re-used where
1334        possible. This sometimes gives as much as 30% improvement.
1335    
1336    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1337        call.
1338    
1339    14. As documented, (*COMMIT) is now confined to within a positive assertion.
1340    
1341    15. It is now possible to link pcretest with libedit as an alternative to
1342        libreadline.
1343    
1344    16. (*COMMIT) control verb is now supported by the JIT compiler.
1345    
1346    17. The Unicode data tables have been updated to Unicode 6.1.0.
1347    
1348    18. Added --file-list option to pcregrep.
1349    
1350    19. Added binary file support to pcregrep, including the -a, --binary-files,
1351        -I, and --text options.
1352    
1353    20. The madvise function is renamed for posix_madvise for QNX compatibility
1354        reasons. Fixed by Giuseppe D'Angelo.
1355    
1356    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1357    
1358    22. Changed the option for creating symbolic links for 16-bit man pages from
1359        -s to -sf so that re-installing does not cause issues.
1360    
1361    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1362    
1363    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1364        matches in certain environments (the workspace was not being correctly
1365        retained). Also added to pcre_dfa_exec() a simple plausibility check on
1366        some of the workspace data at the beginning of a restart.
1367    
1368    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1369        was not doing so when it should - probably a typo introduced by SVN 528
1370        (change 8.10/14).
1371    
1372    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1373        \w+ when the character tables indicated that \x{c4} was a word character.
1374        There were several related cases, all because the tests for doing a table
1375        lookup were testing for characters less than 127 instead of 255.
1376    
1377    27. If a pattern contains capturing parentheses that are not used in a match,
1378        their slots in the ovector are set to -1. For those that are higher than
1379        any matched groups, this happens at the end of processing. In the case when
1380        there were back references that the ovector was too small to contain
1381        (causing temporary malloc'd memory to be used during matching), and the
1382        highest capturing number was not used, memory off the end of the ovector
1383        was incorrectly being set to -1. (It was using the size of the temporary
1384        memory instead of the true size.)
1385    
1386    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1387        ovector size, it uses memory at the end of the block that it has got.
1388    
1389    29. Check for an overlong MARK name and give an error at compile time. The
1390        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1391    
1392    30. JIT compiler update.
1393    
1394    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1395        Rill for the patch.
1396    
1397    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1398    
1399    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1400    
1401    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1402        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1403        was enabled.
1404    
1405    35. Improve JIT code generation for greedy plus quantifier.
1406    
1407    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1408        1 to "aa" instead of to an empty string. The bug affected repeated groups
1409        that could potentially match an empty string.
1410    
1411    37. Optimizing single character iterators in JIT.
1412    
1413    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1414        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1415        codepoints that are too big for the mode are faulted, and in a UTF mode,
1416        disallowed codepoints are also faulted.
1417    
1418    39. If PCRE was compiled with UTF support, in three places in the DFA
1419        matcher there was code that should only have been obeyed in UTF mode, but
1420        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1421        processing when bytes with values greater than 127 were present. In 16-bit
1422        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1423        both cases the values are those that cannot be the first data item in a UTF
1424        character. The three items that might have provoked this were recursions,
1425        possessively repeated groups, and atomic groups.
1426    
1427    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1428        and pcregrep, because some OS require shared objects to be explicitly
1429        passed to ld, causing the link step to fail if they are not.
1430    
1431    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1432        mode, patterns that started with \h* or \R* might be incorrectly matched.
1433    
1434    
1435    Version 8.30 04-February-2012
1436    -----------------------------
1437    
1438    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
1439        name is defined in ctype.h.
1440    
1441    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
1442        only in quite long subpatterns.
1443    
1444    3.  Removed the function pcre_info(), which has been obsolete and deprecated
1445        since it was replaced by pcre_fullinfo() in February 2000.
1446    
1447    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
1448        match a (*MARK), and the match failed at the start of the subject, a
1449        reference to memory before the start of the subject could occur. This bug
1450        was introduced by fix 17 of release 8.21.
1451    
1452    5.  A reference to an unset group with zero minimum repetition was giving
1453        totally wrong answers (in non-JavaScript-compatibility mode). For example,
1454        /(another)?(\1?)test/ matched against "hello world test". This bug was
1455        introduced in release 8.13.
1456    
1457    6.  Add support for 16-bit character strings (a large amount of work involving
1458        many changes and refactorings).
1459    
1460    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1461        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1462        from a file.
1463    
1464    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1465        rounding is not applied in this particular case).
1466    
1467    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1468        if they appear, or are escaped, in patterns.
1469    
1470    10. Get rid of a number of -Wunused-but-set-variable warnings.
1471    
1472    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1473        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1474        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1475        also returns the mark "x". This bug applied to capturing parentheses,
1476        non-capturing parentheses, and atomic parentheses. It also applied to some
1477        assertions.
1478    
1479    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1480        information out of configure.ac instead of relying on pcre.h.generic, which
1481        is not stored in the repository.
1482    
1483    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1484        -lreadline.
1485    
1486    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1487    
1488    15. Applied Graycode's patch to put the top-level frame on the stack rather
1489        than the heap when not using the stack for recursion. This gives a
1490        performance improvement in many cases when recursion is not deep.
1491    
1492    16. Experimental code added to "pcretest -C" to output the stack frame size.
1493    
1494    
1495    Version 8.21 12-Dec-2011
1496    ------------------------
1497    
1498    1.  Updating the JIT compiler.
1499    
1500    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1501        are added as well.
1502    
1503    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1504        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1505        calling _pcre_jit_exec. Some extra comments are added.
1506    
1507    4.  (*MARK) settings inside atomic groups that do not contain any capturing
1508        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1509        was introduced by change 18 for 8.20.
1510    
1511    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1512        ECMA-262 standard.
1513    
1514    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1515        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1516        This bug was probably introduced by change 9 of 8.13.
1517    
1518    7.  While fixing 6 above, I noticed that a number of other items were being
1519        incorrectly rejected as "not fixed length". This arose partly because newer
1520        opcodes had not been added to the fixed-length checking code. I have (a)
1521        corrected the bug and added tests for these items, and (b) arranged for an
1522        error to occur if an unknown opcode is encountered while checking for fixed
1523        length instead of just assuming "not fixed length". The items that were
1524        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1525        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1526        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1527    
1528    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1529        being incorrectly compiled and would have given unpredicatble results.
1530    
1531    9.  A possessively repeated subpattern with minimum repeat count greater than
1532        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1533        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1534        the first (A) could occur when it should not.
1535    
1536    10. Add a cast and remove a redundant test from the code.
1537    
1538    11. JIT should use pcre_malloc/pcre_free for allocation.
1539    
1540    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1541        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1542        is anything other than /usr, -L is still shown).
1543    
1544    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1545    
1546    14. Perl does not support \N without a following name in a [] class; PCRE now
1547        also gives an error.
1548    
1549    15. If a forward reference was repeated with an upper limit of around 2000,
1550        it caused the error "internal error: overran compiling workspace". The
1551        maximum number of forward references (including repeats) was limited by the
1552        internal workspace, and dependent on the LINK_SIZE. The code has been
1553        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1554        the default depends on LINK_SIZE. There is a new upper limit (for safety)
1555        of around 200,000 forward references. While doing this, I also speeded up
1556        the filling in of repeated forward references.
1557    
1558    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1559        incorrectly expecting the subject to contain another "a" after the start.
1560    
1561    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1562        in the match, the SKIP should be ignored. This was not happening; instead
1563        the SKIP was being treated as NOMATCH. For patterns such as
1564        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1565        tested.
1566    
1567    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1568        now much more compatible with Perl, in particular in cases where the result
1569        is a non-match for a non-anchored pattern. For example, if
1570        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1571        "m", where previously it did not return a name. A side effect of this
1572        change is that for partial matches, the last encountered mark name is
1573        returned, as for non matches. A number of tests that were previously not
1574        Perl-compatible have been moved into the Perl-compatible test files. The
1575        refactoring has had the pleasing side effect of removing one argument from
1576        the match() function, thus reducing its stack requirements.
1577    
1578    19. If the /S+ option was used in pcretest to study a pattern using JIT,
1579        subsequent uses of /S (without +) incorrectly behaved like /S+.
1580    
1581    21. Retrieve executable code size support for the JIT compiler and fixing
1582        some warnings.
1583    
1584    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1585        not work when the shorter character appeared right at the end of the
1586        subject string.
1587    
1588    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1589        systems.
1590    
1591    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1592        output it when the /M option is used in pcretest.
1593    
1594    25. The CheckMan script was not being included in the distribution. Also, added
1595        an explicit "perl" to run Perl scripts from the PrepareRelease script
1596        because this is reportedly needed in Windows.
1597    
1598    26. If study data was being save in a file and studying had not found a set of
1599        "starts with" bytes for the pattern, the data written to the file (though
1600        never used) was taken from uninitialized memory and so caused valgrind to
1601        complain.
1602    
1603    27. Updated RunTest.bat as provided by Sheri Pierce.
1604    
1605    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1606    
1607    29. Computation of memory usage for the table of capturing group names was
1608        giving an unnecessarily large value.
1609    
1610    
1611    Version 8.20 21-Oct-2011
1612    ------------------------
1613    
1614    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1615        a POSIX class. After further experiments with Perl, which convinced me that
1616        Perl has bugs and confusions, a closing square bracket is no longer allowed
1617        in a POSIX name. This bug also affected patterns with classes that started
1618        with full stops.
1619    
1620    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1621        captured substring, but while checking the failing first alternative,
1622        substring 1 is temporarily captured. If the output vector supplied to
1623        pcre_exec() was not big enough for this capture, the yield of the function
1624        was still zero ("insufficient space for captured substrings"). This cannot
1625        be totally fixed without adding another stack variable, which seems a lot
1626        of expense for a edge case. However, I have improved the situation in cases
1627        such as /(a)(b)x|abc/ matched against "abc", where the return code
1628        indicates that fewer than the maximum number of slots in the ovector have
1629        been set.
1630    
1631    3.  Related to (2) above: when there are more back references in a pattern than
1632        slots in the output vector, pcre_exec() uses temporary memory during
1633        matching, and copies in the captures as far as possible afterwards. It was
1634        using the entire output vector, but this conflicts with the specification
1635        that only 2/3 is used for passing back captured substrings. Now it uses
1636        only the first 2/3, for compatibility. This is, of course, another edge
1637        case.
1638    
1639    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
1640        main code base, and can be used by building with --enable-jit. When this is
1641        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1642        runtime --no-jit option is given.
1643    
1644    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
1645        ovector, the return from the function was zero, implying that there were
1646        other matches that did not fit. The correct "exactly full" value is now
1647        returned.
1648    
1649    6.  If a subpattern that was called recursively or as a subroutine contained
1650        (*PRUNE) or any other control that caused it to give a non-standard return,
1651        invalid errors such as "Error -26 (nested recursion at the same subject
1652        position)" or even infinite loops could occur.
1653    
1654    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1655        computing the minimum length on reaching *ACCEPT, and so ended up with the
1656        wrong value of 1 rather than 0. Further investigation indicates that
1657        computing a minimum subject length in the presence of *ACCEPT is difficult
1658        (think back references, subroutine calls), and so I have changed the code
1659        so that no minimum is registered for a pattern that contains *ACCEPT.
1660    
1661    8.  If (*THEN) was present in the first (true) branch of a conditional group,
1662        it was not handled as intended. [But see 16 below.]
1663    
1664    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1665        Sheri Pierce.
1666    
1667    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1668        the first byte in a match must be "a".
1669    
1670    11. Change 17 for 8.13 increased the recursion depth for patterns like
1671        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1672        pattern contains any instances of (*THEN). If it does not, the old
1673        optimizations are restored. It would be nice to do this on a per-group
1674        basis, but at the moment that is not feasible.
1675    
1676    12. In some environments, the output of pcretest -C is CRLF terminated. This
1677        broke RunTest's code that checks for the link size. A single white space
1678        character after the value is now allowed for.
1679    
1680    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1681        For "fr", it uses the Windows-specific input and output files.
1682    
1683    14. If (*THEN) appeared in a group that was called recursively or as a
1684        subroutine, it did not work as intended. [But see next item.]
1685    
1686    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1687        pattern fragments (but not containing any | characters). If A and B are
1688        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1689        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1690        D. In other words, Perl considers parentheses that do not contain any |
1691        characters to be part of a surrounding alternative, whereas PCRE was
1692        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1693        differently. PCRE now behaves in the same way as Perl, except in the case
1694        of subroutine/recursion calls such as (?1) which have in any case always
1695        been different (but PCRE had them first :-).
1696    
1697    16. Related to 15 above: Perl does not treat the | in a conditional group as
1698        creating alternatives. Such a group is treated in the same way as an
1699        ordinary group without any | characters when processing (*THEN). PCRE has
1700        been changed to match Perl's behaviour.
1701    
1702    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1703        RunGrepTest script failed.
1704    
1705    18. Change 22 for version 13 caused atomic groups to use more stack. This is
1706        inevitable for groups that contain captures, but it can lead to a lot of
1707        stack use in large patterns. The old behaviour has been restored for atomic
1708        groups that do not contain any capturing parentheses.
1709    
1710    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1711        suppress the check for a minimum subject length at run time. (If it was
1712        given to pcre_exec() or pcre_dfa_exec() it did work.)
1713    
1714    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1715        fail to work when decoding hex characters in data strings in EBCDIC
1716        environments.
1717    
1718    21. It appears that in at least one Mac OS environment, the isxdigit() function
1719        is implemented as a macro that evaluates to its argument more than once,
1720        contravening the C 90 Standard (I haven't checked a later standard). There
1721        was an instance in pcretest which caused it to go wrong when processing
1722        \x{...} escapes in subject strings. The has been rewritten to avoid using
1723        things like p++ in the argument of isxdigit().
1724    
1725    
1726    Version 8.13 16-Aug-2011
1727    ------------------------
1728    
1729    1.  The Unicode data tables have been updated to Unicode 6.0.0.
1730    
1731    2.  Two minor typos in pcre_internal.h have been fixed.
1732    
1733    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1734        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1735        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1736    
1737    4.  There were a number of related bugs in the code for matching backrefences
1738        caselessly in UTF-8 mode when codes for the characters concerned were
1739        different numbers of bytes. For example, U+023A and U+2C65 are an upper
1740        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1741        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1742        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1743        2-byte code at the end of the subject (it thought there wasn't enough data
1744        left).
1745    
1746    5.  Comprehensive information about what went wrong is now returned by
1747        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1748        as the output vector has at least 2 elements. The offset of the start of
1749        the failing character and a reason code are placed in the vector.
1750    
1751    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
1752        now returned is for the first byte of the failing character, instead of the
1753        last byte inspected. This is an incompatible change, but I hope it is small
1754        enough not to be a problem. It makes the returned offset consistent with
1755        pcre_exec() and pcre_dfa_exec().
1756    
1757    7.  pcretest now gives a text phrase as well as the error number when
1758        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1759        failure, the offset and reason code are output.
1760    
1761    8.  When \R was used with a maximizing quantifier it failed to skip backwards
1762        over a \r\n pair if the subsequent match failed. Instead, it just skipped
1763        back over a single character (\n). This seems wrong (because it treated the
1764        two characters as a single entity when going forwards), conflicts with the
1765        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1766        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1767        has been changed.
1768    
1769    9.  Some internal refactoring has changed the processing so that the handling
1770        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1771        time (the PCRE_DOTALL option was changed this way some time ago: version
1772        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1773        which was always a bit of a fudge. It also means that there is one less
1774        argument for the match() function, which reduces its stack requirements
1775        slightly. This change also fixes an incompatibility with Perl: the pattern
1776        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1777    
1778    10. More internal refactoring has drastically reduced the number of recursive
1779        calls to match() for possessively repeated groups such as (abc)++ when
1780        using pcre_exec().
1781    
1782    11. While implementing 10, a number of bugs in the handling of groups were
1783        discovered and fixed:
1784    
1785        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1786        (a|)*(?1) gave a compile-time internal error.
1787        ((a|)+)+  did not notice that the outer group could match an empty string.
1788        (^a|^)+   was not marked as anchored.
1789        (.*a|.*)+ was not marked as matching at start or after a newline.
1790    
1791    12. Yet more internal refactoring has removed another argument from the match()
1792        function. Special calls to this function are now indicated by setting a
1793        value in a variable in the "match data" data block.
1794    
1795    13. Be more explicit in pcre_study() instead of relying on "default" for
1796        opcodes that mean there is no starting character; this means that when new
1797        ones are added and accidentally left out of pcre_study(), testing should
1798        pick them up.
1799    
1800    14. The -s option of pcretest has been documented for ages as being an old
1801        synonym of -m (show memory usage). I have changed it to mean "force study
1802        for every regex", that is, assume /S for every regex. This is similar to -i
1803        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1804        using it. It makes it easier to run collections of tests with and without
1805        study enabled, and thereby test pcre_study() more easily. All the standard
1806        tests are now run with and without -s (but some patterns can be marked as
1807        "never study" - see 20 below).
1808    
1809    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1810        restoration of the capturing data to the outer values was not happening
1811        correctly.
1812    
1813    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1814        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1815        pattern had matched an empty string, and so incorrectly returned a no
1816        match.
1817    
1818    17. There was optimizing code for the last branch of non-capturing parentheses,
1819        and also for the obeyed branch of a conditional subexpression, which used
1820        tail recursion to cut down on stack usage. Unfortunately, now that there is
1821        the possibility of (*THEN) occurring in these branches, tail recursion is
1822        no longer possible because the return has to be checked for (*THEN). These
1823        two optimizations have therefore been removed. [But see 8.20/11 above.]
1824    
1825    18. If a pattern containing \R was studied, it was assumed that \R always
1826        matched two bytes, thus causing the minimum subject length to be
1827        incorrectly computed because \R can also match just one byte.
1828    
1829    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1830        was incorrectly computed.
1831    
1832    20. If /S is present twice on a test pattern in pcretest input, it now
1833        *disables* studying, thereby overriding the use of -s on the command line
1834        (see 14 above). This is necessary for one or two tests to keep the output
1835        identical in both cases.
1836    
1837    21. When (*ACCEPT) was used in an assertion that matched an empty string and
1838        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1839    
1840    22. When an atomic group that contained a capturing parenthesis was
1841        successfully matched, but the branch in which it appeared failed, the
1842        capturing was not being forgotten if a higher numbered group was later
1843        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1844        group 1 to "a", when in fact it should be unset. This applied to multi-
1845        branched capturing and non-capturing groups, repeated or not, and also to
1846        positive assertions (capturing in negative assertions does not happen
1847        in PCRE) and also to nested atomic groups.
1848    
1849    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1850        subject after a captured substring, to make it easier to tell which of a
1851        number of identical substrings has been captured.
1852    
1853    24. The way atomic groups are processed by pcre_exec() has been changed so that
1854        if they are repeated, backtracking one repetition now resets captured
1855        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1856        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1857        "aaa". Previously, it would have been "a". As part of this code
1858        refactoring, the way recursive calls are handled has also been changed.
1859    
1860    25. If an assertion condition captured any substrings, they were not passed
1861        back unless some other capturing happened later. For example, if
1862        (?(?=(a))a) was matched against "a", no capturing was returned.
1863    
1864    26. When studying a pattern that contained subroutine calls or assertions,
1865        the code for finding the minimum length of a possible match was handling
1866        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1867        group 1 called group 2 while simultaneously a separate group 2 called group
1868        1). A stack overflow occurred in this case. I have fixed this by limiting
1869        the recursion depth to 10.
1870    
1871    27. Updated RunTest.bat in the distribution to the version supplied by Tom
1872        Fortmann. This supports explicit test numbers on the command line, and has
1873        argument validation and error reporting.
1874    
1875    28. An instance of \X with an unlimited repeat could fail if at any point the
1876        first character it looked at was a mark character.
1877    
1878    29. Some minor code refactoring concerning Unicode properties and scripts
1879        should reduce the stack requirement of match() slightly.
1880    
1881    30. Added the '=' option to pcretest to check the setting of unused capturing
1882        slots at the end of the pattern, which are documented as being -1, but are
1883        not included in the return count.
1884    
1885    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1886        compiled something random. Now it gives a compile-time error (as does
1887        Perl).
1888    
1889    32. A *MARK encountered during the processing of a positive assertion is now
1890        recorded and passed back (compatible with Perl).
1891    
1892    33. If --only-matching or --colour was set on a pcregrep call whose pattern
1893        had alternative anchored branches, the search for a second match in a line
1894        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1895        matched the line "0102" twice. The same bug affected patterns that started
1896        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1897        twice.
1898    
1899    34. Previously, PCRE did not allow quantification of assertions. However, Perl
1900        does, and because of capturing effects, quantifying parenthesized
1901        assertions may at times be useful. Quantifiers are now allowed for
1902        parenthesized assertions.
1903    
1904    35. A minor code tidy in pcre_compile() when checking options for \R usage.
1905    
1906    36. \g was being checked for fancy things in a character class, when it should
1907        just be a literal "g".
1908    
1909    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1910        appearance of a nested POSIX class supersedes an apparent external class.
1911        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1912        unescaped square brackets may also appear as part of class names. For
1913        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1914        more like Perl. (But see 8.20/1 above.)
1915    
1916    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1917        was because it thought it was \N{name}, which is not supported).
1918    
1919    39. Add minix to OS list not supporting the -S option in pcretest.
1920    
1921    40. PCRE tries to detect cases of infinite recursion at compile time, but it
1922        cannot analyze patterns in sufficient detail to catch mutual recursions
1923        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1924        subgroup is called recursively as a subpattern for a second time at the
1925        same position in the subject string. In previous releases this might have
1926        been caught by the recursion limit, or it might have run out of stack.
1927    
1928    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1929        happen only once. PCRE was, however incorrectly giving a compile time error
1930        "recursive call could loop indefinitely" because it cannot analyze the
1931        pattern in sufficient detail. The compile time test no longer happens when
1932        PCRE is compiling a conditional subpattern, but actual runaway loops are
1933        now caught at runtime (see 40 above).
1934    
1935    42. It seems that Perl allows any characters other than a closing parenthesis
1936        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1937        has been changed to be the same.
1938    
1939    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1940        as not to get warnings when autogen.sh is called. Also changed
1941        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1942    
1943    44. To help people who use pcregrep to scan files containing exceedingly long
1944        lines, the following changes have been made:
1945    
1946        (a) The default value of the buffer size parameter has been increased from
1947            8K to 20K. (The actual buffer used is three times this size.)
1948    
1949        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1950            PCRE is built.
1951    
1952        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1953            to be set at run time.
1954    
1955        (d) Numerical values in pcregrep options can be followed by K or M, for
1956            example --buffer-size=50K.
1957    
1958        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1959            given and the return code is set to 2.
1960    
1961    45. Add a pointer to the latest mark to the callout data block.
1962    
1963    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1964        partial match of an empty string instead of no match. This was specific to
1965        the use of ".".
1966    
1967    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1968        complete match instead of a partial match. This bug was dependent on both
1969        the PCRE_UTF8 and PCRE_DOTALL options being set.
1970    
1971    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1972        starting byte set, because \b was not being ignored.
1973    
1974    
1975    Version 8.12 15-Jan-2011
1976    ------------------------
1977    
1978    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1979        checks for such things as part of the documentation building process.
1980    
1981    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1982        --match-limit and --recursion-limit options (added for 8.11). In
1983        particular, this made one of the standard tests fail. (The integer value
1984        went into the wrong half of a long int.)
1985    
1986    3.  If the --colour option was given to pcregrep with -v (invert match), it
1987        did strange things, either producing crazy output, or crashing. It should,
1988        of course, ignore a request for colour when reporting lines that do not
1989        match.
1990    
1991    4.  Another pcregrep bug caused similar problems if --colour was specified with
1992        -M (multiline) and the pattern match finished with a line ending.
1993    
1994    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1995        matched in multiline mode, the following line was shown as part of the
1996        match. This seems wrong, so I have changed it.
1997    
1998    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1999        the check for further matches in the same line (so they could be coloured)
2000        to overrun the end of the current line. If another match was found, it was
2001        incorrectly shown (and then shown again when found in the next line).
2002    
2003    7.  If pcregrep was compiled under Windows, there was a reference to the
2004        function pcregrep_exit() before it was defined. I am assuming this was
2005        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
2006        reported by a user. I've moved the definition above the reference.
2007    
2008    
2009    Version 8.11 10-Dec-2010
2010    ------------------------
2011    
2012    1.  (*THEN) was not working properly if there were untried alternatives prior
2013        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
2014        backtracked to try for "b" instead of moving to the next alternative branch
2015        at the same level (in this case, to look for "c"). The Perl documentation
2016        is clear that when (*THEN) is backtracked onto, it goes to the "next
2017        alternative in the innermost enclosing group".
2018    
2019    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
2020        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
2021        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
2022        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
2023        (*THEN).
2024    
2025    3.  If \s appeared in a character class, it removed the VT character from
2026        the class, even if it had been included by some previous item, for example
2027        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
2028        of \s, but is part of the POSIX "space" class.)
2029    
2030    4.  A partial match never returns an empty string (because you can always
2031        match an empty string at the end of the subject); however the checking for
2032        an empty string was starting at the "start of match" point. This has been
2033        changed to the "earliest inspected character" point, because the returned
2034        data for a partial match starts at this character. This means that, for
2035        example, /(?<=abc)def/ gives a partial match for the subject "abc"
2036        (previously it gave "no match").
2037    
2038    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
2039        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
2040        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
2041        has an implication that the given string is incomplete (because a partial
2042        match is preferred over a full match). For this reason, these items now
2043        give a partial match in this situation. [Aside: previously, the one case
2044        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
2045        match rather than a full match, which was wrong by the old rules, but is
2046        now correct.]
2047    
2048    6.  There was a bug in the handling of #-introduced comments, recognized when
2049        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
2050        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
2051        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
2052        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
2053        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
2054        places in pcre_compile().
2055    
2056    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
2057        comments when looking ahead for named forward references to subpatterns,
2058        the only newline sequence it recognized was NL. It now handles newlines
2059        according to the set newline convention.
2060    
2061    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
2062        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
2063        cater for a lack of strerror(). These oversights have been fixed.
2064    
2065    9.  Added --match-limit and --recursion-limit to pcregrep.
2066    
2067    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
2068    
2069    11. When the -o option was used, pcregrep was setting a return code of 1, even
2070        when matches were found, and --line-buffered was not being honoured.
2071    
2072    12. Added an optional parentheses number to the -o and --only-matching options
2073        of pcregrep.
2074    
2075    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
2076        can match an empty string. The code to do it in pcretest and pcredemo
2077        needed fixing:
2078    
2079        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
2080            only one byte after an empty string match just before CRLF (this case
2081            just got forgotten; "any" and "anycrlf" were OK).
2082    
2083        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
2084            mode when an empty string match preceded an ASCII character followed by
2085            a non-ASCII character. (The code for advancing by one character rather
2086            than one byte was nonsense.)
2087    
2088        (c) The pcredemo.c sample program did not have any code at all to handle
2089            the cases when CRLF is a valid newline sequence.
2090    
2091    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2092        as a starting offset was within the subject string. There is now a new
2093        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2094        negative or greater than the length of the string. In order to test this,
2095        pcretest is extended to allow the setting of negative starting offsets.
2096    
2097    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2098        starting offset points to the beginning of a UTF-8 character was
2099        unnecessarily clumsy. I tidied it up.
2100    
2101    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2102        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2103    
2104    17. Nobody had reported that the --include_dir option, which was added in
2105        release 7.7 should have been called --include-dir (hyphen, not underscore)
2106        for compatibility with GNU grep. I have changed it to --include-dir, but
2107        left --include_dir as an undocumented synonym, and the same for
2108        --exclude-dir, though that is not available in GNU grep, at least as of
2109        release 2.5.4.
2110    
2111    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2112        characters from a string of bytes) have been redefined so as not to use
2113        loops, in order to improve performance in some environments. At the same
2114        time, I abstracted some of the common code into auxiliary macros to save
2115        repetition (this should not affect the compiled code).
2116    
2117    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2118        compile-time error is now given if \c is not followed by an ASCII
2119        character, that is, a byte less than 128. (In EBCDIC mode, the code is
2120        different, and any byte value is allowed.)
2121    
2122    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2123        START_OPTIMIZE option, which is now allowed at compile time - but just
2124        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2125        to pcregrep and other applications that have no direct access to PCRE
2126        options. The new /Y option in pcretest sets this option when calling
2127        pcre_compile().
2128    
2129    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2130        back references. Groups containing recursive back references were forced to
2131        be atomic by that change, but in the case of named groups, the amount of
2132        memory required was incorrectly computed, leading to "Failed: internal
2133        error: code overflow". This has been fixed.
2134    
2135    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2136        pcretest.c, to avoid build problems in some Borland environments.
2137    
2138    
2139    Version 8.10 25-Jun-2010
2140    ------------------------
2141    
2142    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2143        THEN.
2144    
2145    2.  (*ACCEPT) was not working when inside an atomic group.
2146    
2147    3.  Inside a character class, \B is treated as a literal by default, but
2148        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2149        causes the error). The code is unchanged, but I tidied the documentation.
2150    
2151    4.  Inside a character class, PCRE always treated \R and \X as literals,
2152        whereas Perl faults them if its -w option is set. I have changed PCRE so
2153        that it faults them when PCRE_EXTRA is set.
2154    
2155    5.  Added support for \N, which always matches any character other than
2156        newline. (It is the same as "." when PCRE_DOTALL is not set.)
2157    
2158    6.  When compiling pcregrep with newer versions of gcc which may have
2159        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2160        declared with attribute warn_unused_result" were given. Just casting the
2161        result to (void) does not stop the warnings; a more elaborate fudge is
2162        needed. I've used a macro to implement this.
2163    
2164    7.  Minor change to pcretest.c to avoid a compiler warning.
2165    
2166    8.  Added four artifical Unicode properties to help with an option to make
2167        \s etc use properties (see next item). The new properties are: Xan
2168        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2169    
2170    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2171        use Unicode properties. (*UCP) at the start of a pattern can be used to set
2172        this option. Modified pcretest to add /W to test this facility. Added
2173        REG_UCP to make it available via the POSIX interface.
2174    
2175    10. Added --line-buffered to pcregrep.
2176    
2177    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2178        studied, and the match started with a letter with a code point greater than
2179        127 whose first byte was different to the first byte of the other case of
2180        the letter, the other case of this starting letter was not recognized
2181        (#976).
2182    
2183    12. If a pattern that was studied started with a repeated Unicode property
2184        test, for example, \p{Nd}+, there was the theoretical possibility of
2185        setting up an incorrect bitmap of starting bytes, but fortunately it could
2186        not have actually happened in practice until change 8 above was made (it
2187        added property types that matched character-matching opcodes).
2188    
2189    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2190        possible starting bytes for non-anchored patterns.
2191    
2192    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2193        \R, and also a number of cases that involve Unicode properties, both
2194        explicit and implicit when PCRE_UCP is set.
2195    
2196    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2197        input, it could crash or give wrong results if characters with values
2198        greater than 0xc0 were present in the subject string. (Detail: it assumed
2199        UTF-8 input when processing these items.)
2200    
2201    16. Added a lot of (int) casts to avoid compiler warnings in systems where
2202        size_t is 64-bit (#991).
2203    
2204    17. Added a check for running out of memory when PCRE is compiled with
2205        --disable-stack-for-recursion (#990).
2206    
2207    18. If the last data line in a file for pcretest does not have a newline on
2208        the end, a newline was missing in the output.
2209    
2210    19. The default pcre_chartables.c file recognizes only ASCII characters (values
2211        less than 128) in its various bitmaps. However, there is a facility for
2212        generating tables according to the current locale when PCRE is compiled. It
2213        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2214        characters, are recognized by isspace() and therefore were getting set in
2215        these tables, and indeed these tables seem to approximate to ISO 8859. This
2216        caused a problem in UTF-8 mode when pcre_study() was used to create a list
2217        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2218        which of course cannot start UTF-8 characters. I have changed the code so
2219        that only real ASCII characters (less than 128) and the correct starting
2220        bytes for UTF-8 encodings are set for characters greater than 127 when in
2221        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2222        altogether.)
2223    
2224    20. Added the /T option to pcretest so as to be able to run tests with non-
2225        standard character tables, thus making it possible to include the tests
2226        used for 19 above in the standard set of tests.
2227    
2228    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2229        reference to a subpattern the other side of a comment that contains an
2230        opening parenthesis caused either an internal compiling error, or a
2231        reference to the wrong subpattern.
2232    
2233    
2234    Version 8.02 19-Mar-2010
2235    ------------------------
2236    
2237    1.  The Unicode data tables have been updated to Unicode 5.2.0.
2238    
2239    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
2240        configured.
2241    
2242    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
2243        original author of that file, following a query about its status.
2244    
2245    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
2246        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2247    
2248    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2249        quantifier applied to a forward-referencing subroutine call, could compile
2250        incorrect code or give the error "internal error: previously-checked
2251        referenced subpattern not found".
2252    
2253    6.  Both MS Visual Studio and Symbian OS have problems with initializing
2254        variables to point to external functions. For these systems, therefore,
2255        pcre_malloc etc. are now initialized to local functions that call the
2256        relevant global functions.
2257    
2258    7.  There were two entries missing in the vectors called coptable and poptable
2259        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2260        I've fixed the data, and added a kludgy way of testing at compile time that
2261        the lengths are correct (equal to the number of opcodes).
2262    
2263    8.  Following on from 7, I added a similar kludge to check the length of the
2264        eint vector in pcreposix.c.
2265    
2266    9.  Error texts for pcre_compile() are held as one long string to avoid too
2267        much relocation at load time. To find a text, the string is searched,
2268        counting zeros. There was no check for running off the end of the string,
2269        which could happen if a new error number was added without updating the
2270        string.
2271    
2272    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2273    
2274    11. \K was not working if it appeared in an atomic group or in a group that
2275        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2276        \K is "not well defined" if used in an assertion. PCRE now accepts it if
2277        the assertion is positive, but not if it is negative.
2278    
2279    12. Change 11 fortuitously reduced the size of the stack frame used in the
2280        "match()" function of pcre_exec.c by one pointer. Forthcoming
2281        implementation of support for (*MARK) will need an extra pointer on the
2282        stack; I have reserved it now, so that the stack frame size does not
2283        decrease.
2284    
2285    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2286        item in branch that calls a recursion is a subroutine call - as in the
2287        second branch in the above example - was incorrectly given the compile-
2288        time error "recursive call could loop indefinitely" because pcre_compile()
2289        was not correctly checking the subroutine for matching a non-empty string.
2290    
2291    14. The checks for overrunning compiling workspace could trigger after an
2292        overrun had occurred. This is a "should never occur" error, but it can be
2293        triggered by pathological patterns such as hundreds of nested parentheses.
2294        The checks now trigger 100 bytes before the end of the workspace.
2295    
2296    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2297    
2298    
2299    Version 8.01 19-Jan-2010
2300    ------------------------
2301    
2302    1.  If a pattern contained a conditional subpattern with only one branch (in
2303        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2304        computed the wrong minimum data length (which is of course zero for such
2305        subpatterns). This could cause incorrect "no match" results.
2306    
2307    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2308        the pattern is reset in the first branch, pcre_compile() failed with
2309        "internal error: code overflow at offset...". This happened only when
2310        the reset was to the original external option setting. (An optimization
2311        abstracts leading options settings into an external setting, which was the
2312        cause of this.)
2313    
2314    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2315        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2316        assertion pattern did not match (meaning that the assertion was true), it
2317        was incorrectly treated as false if the SKIP had been reached during the
2318        matching. This also applied to assertions used as conditions.
2319    
2320    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
2321        assertion subpattern, including such a pattern used as a condition,
2322        unpredictable results occurred, instead of the error return
2323        PCRE_ERROR_DFA_UITEM.
2324    
2325    5.  The C++ GlobalReplace function was not working like Perl for the special
2326        situation when an empty string is matched. It now does the fancy magic
2327        stuff that is necessary.
2328    
2329    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2330        removed. (These were left over from very, very early versions of PCRE.)
2331    
2332    7.  Some cosmetic changes to the code to make life easier when compiling it
2333        as part of something else:
2334    
2335        (a) Change DEBUG to PCRE_DEBUG.
2336    
2337        (b) In pcre_compile(), rename the member of the "branch_chain" structure
2338            called "current" as "current_branch", to prevent a collision with the
2339            Linux macro when compiled as a kernel module.
2340    
2341        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2342            prevent a collision with the Linux macro when compiled as a kernel
2343            module.
2344    
2345    8.  In pcre_compile() there are some checks for integer overflows that used to
2346        cast potentially large values to (double). This has been changed to that
2347        when building, a check for int64_t is made, and if it is found, it is used
2348        instead, thus avoiding the use of floating point arithmetic. (There is no
2349        other use of FP in PCRE.) If int64_t is not found, the fallback is to
2350        double.
2351    
2352    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
2353        2005 (difference between two addresses compared to an unsigned value).
2354    
2355    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2356        custom one, because of the following reported problem in Windows:
2357    
2358          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2359              under Win32.
2360          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2361              therefore missing the function definition.
2362          - The compiler thus generates a "C" signature for the test function.
2363          - The linker fails to find the "C" function.
2364          - PCRE fails to configure if asked to do so against libbz2.
2365    
2366    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2367        messages were output:
2368    
2369          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2370          rerunning libtoolize, to keep the correct libtool macros in-tree.
2371          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2372    
2373        I have done both of these things.
2374    
2375    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2376        most of the time, it *can* run out if it is given a pattern that contains a
2377        runaway infinite recursion. I updated the discussion in the pcrestack man
2378        page.
2379    
2380    13. Now that we have gone to the x.xx style of version numbers, the minor
2381        version may start with zero. Using 08 or 09 is a bad idea because users
2382        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2383        interpreted as invalid octal numbers. I've updated the previous comment in
2384        configure.ac, and also added a check that gives an error if 08 or 09 are
2385        used.
2386    
2387    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2388        causing partial matching to fail when the end of the subject matched \W
2389        in a UTF-8 pattern where \W was quantified with a minimum of 3.
2390    
2391    15. There were some discrepancies between the declarations in pcre_internal.h
2392        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2393        their definitions. The declarations used "const uschar *" and the
2394        definitions used USPTR. Even though USPTR is normally defined as "const
2395        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2396        reported that: "This difference in casting confuses some C++ compilers, for
2397        example, SunCC recognizes above declarations as different functions and
2398        generates broken code for hbpcre." I have changed the declarations to use
2399        USPTR.
2400    
2401    16. GNU libtool is named differently on some systems. The autogen.sh script now
2402        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2403        (FreeBSD).
2404    
2405    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2406        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2407        comment: "Figure out how to create a longlong from a string: strtoll and
2408        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2409        instance, but it only takes 2 args instead of 3!"
2410    
2411    18. A subtle bug concerned with back references has been fixed by a change of
2412        specification, with a corresponding code fix. A pattern such as
2413        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2414        refers, was giving matches when it shouldn't. For example, xa=xaaa would
2415        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2416        same bug. Such groups have to be quantified to be useful, or contained
2417        inside another quantified group. (If there's no repetition, the reference
2418        can never match.) The problem arises because, having left the group and
2419        moved on to the rest of the pattern, a later failure that backtracks into
2420        the group uses the captured value from the final iteration of the group
2421        rather than the correct earlier one. I have fixed this in PCRE by forcing
2422        any group that contains a reference to itself to be an atomic group; that
2423        is, there cannot be any backtracking into it once it has completed. This is
2424        similar to recursive and subroutine calls.
2425    
2426    
2427    Version 8.00 19-Oct-09
2428    ----------------------
2429    
2430    1.  The table for translating pcre_compile() error codes into POSIX error codes
2431        was out-of-date, and there was no check on the pcre_compile() error code
2432        being within the table. This could lead to an OK return being given in
2433        error.
2434    
2435    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
2436        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2437        in a Windows environment.
2438    
2439    3.  The pcregrep --count option prints the count for each file even when it is
2440        zero, as does GNU grep. However, pcregrep was also printing all files when
2441        --files-with-matches was added. Now, when both options are given, it prints
2442        counts only for those files that have at least one match. (GNU grep just
2443        prints the file name in this circumstance, but including the count seems
2444        more useful - otherwise, why use --count?) Also ensured that the
2445        combination -clh just lists non-zero counts, with no names.
2446    
2447    4.  The long form of the pcregrep -F option was incorrectly implemented as
2448        --fixed_strings instead of --fixed-strings. This is an incompatible change,
2449        but it seems right to fix it, and I didn't think it was worth preserving
2450        the old behaviour.
2451    
2452    5.  The command line items --regex=pattern and --regexp=pattern were not
2453        recognized by pcregrep, which required --regex pattern or --regexp pattern
2454        (with a space rather than an '='). The man page documented the '=' forms,
2455        which are compatible with GNU grep; these now work.
2456    
2457    6.  No libpcreposix.pc file was created for pkg-config; there was just
2458        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2459    
2460    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2461        when UCP support is not needed, by modifying the Python script that
2462        generates it from Unicode data files. This should not matter if the module
2463        is correctly used as a library, but I received one complaint about 50K of
2464        unwanted data. My guess is that the person linked everything into his
2465        program rather than using a library. Anyway, it does no harm.
2466    
2467    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2468        was a minimum greater than 1 for a wide character in a possessive
2469        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2470        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2471        character. Chaos in the form of incorrect output or a compiling loop could
2472        result.
2473    
2474    9.  The restrictions on what a pattern can contain when partial matching is
2475        requested for pcre_exec() have been removed. All patterns can now be
2476        partially matched by this function. In addition, if there are at least two
2477        slots in the offset vector, the offset of the earliest inspected character
2478        for the match and the offset of the end of the subject are set in them when
2479        PCRE_ERROR_PARTIAL is returned.
2480    
2481    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2482        synonymous with PCRE_PARTIAL, for backwards compatibility, and
2483        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2484        and may be more useful for multi-segment matching.
2485    
2486    11. Partial matching with pcre_exec() is now more intuitive. A partial match
2487        used to be given if ever the end of the subject was reached; now it is
2488        given only if matching could not proceed because another character was
2489        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2490        string "Z", which now yields "no match" instead of "partial match". In the
2491        case of pcre_dfa_exec(), "no match" is given if every matching path for the
2492        final character ended with (*FAIL).
2493    
2494    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2495        if the pattern had a "must contain" character that was already found in the
2496        earlier partial match, unless partial matching was again requested. For
2497        example, with the pattern /dog.(body)?/, the "must contain" character is
2498        "g". If the first part-match was for the string "dog", restarting with
2499        "sbody" failed. This bug has been fixed.
2500    
2501    13. The string returned by pcre_dfa_exec() after a partial match has been
2502        changed so that it starts at the first inspected character rather than the
2503        first character of the match. This makes a difference only if the pattern
2504        starts with a lookbehind assertion or \b or \B (\K is not supported by
2505        pcre_dfa_exec()). It's an incompatible change, but it makes the two
2506        matching functions compatible, and I think it's the right thing to do.
2507    
2508    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2509        so that the demonstration program is easily available in environments where
2510        PCRE has not been installed from source.
2511    
2512    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2513        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2514        library.
2515    
2516    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2517        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2518        is not the first non-POSIX option to be added. Clearly some people find
2519        these options useful.
2520    
2521    17. If a caller to the POSIX matching function regexec() passes a non-zero
2522        value for nmatch with a NULL value for pmatch, the value of
2523        nmatch is forced to zero.
2524    
2525    18. RunGrepTest did not have a test for the availability of the -u option of
2526        the diff command, as RunTest does. It now checks in the same way as
2527        RunTest, and also checks for the -b option.
2528    
2529    19. If an odd number of negated classes containing just a single character
2530        interposed, within parentheses, between a forward reference to a named
2531        subpattern and the definition of the subpattern, compilation crashed with
2532        an internal error, complaining that it could not find the referenced
2533        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2534        [The bug was that it was starting one character too far in when skipping
2535        over the character class, thus treating the ] as data rather than
2536        terminating the class. This meant it could skip too much.]
2537    
2538    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2539        /g option in pcretest when the pattern contains \K, which makes it possible
2540        to have an empty string match not at the start, even when the pattern is
2541        anchored. Updated pcretest and pcredemo to use this option.
2542    
2543    21. If the maximum number of capturing subpatterns in a recursion was greater
2544        than the maximum at the outer level, the higher number was returned, but
2545        with unset values at the outer level. The correct (outer level) value is
2546        now given.
2547    
2548    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2549        PCRE did not set those parentheses (unlike Perl). I have now found a way to
2550        make it do so. The string so far is captured, making this feature
2551        compatible with Perl.
2552    
2553    23. The tests have been re-organized, adding tests 11 and 12, to make it
2554        possible to check the Perl 5.10 features against Perl 5.10.
2555    
2556    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2557        pattern matches a fixed length string. PCRE did not allow this; now it
2558        does. Neither allows recursion.
2559    
2560    25. I finally figured out how to implement a request to provide the minimum
2561        length of subject string that was needed in order to match a given pattern.
2562        (It was back references and recursion that I had previously got hung up
2563        on.) This code has now been added to pcre_study(); it finds a lower bound
2564        to the length of subject needed. It is not necessarily the greatest lower
2565        bound, but using it to avoid searching strings that are too short does give
2566        some useful speed-ups. The value is available to calling programs via
2567        pcre_fullinfo().
2568    
2569    26. While implementing 25, I discovered to my embarrassment that pcretest had
2570        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2571        study optimizations had never been tested with that matching function.
2572        Oops. What is worse, even when it was passed study data, there was a bug in
2573        pcre_dfa_exec() that meant it never actually used it. Double oops. There
2574        were also very few tests of studied patterns with pcre_dfa_exec().
2575    
2576    27. If (?| is used to create subpatterns with duplicate numbers, they are now
2577        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2578        on the other side of the coin, they are no longer allowed to have different
2579        names, because these cannot be distinguished in PCRE, and this has caused
2580        confusion. (This is a difference from Perl.)
2581    
2582    28. When duplicate subpattern names are present (necessarily with different
2583        numbers, as required by 27 above), and a test is made by name in a
2584        conditional pattern, either for a subpattern having been matched, or for
2585        recursion in such a pattern, all the associated numbered subpatterns are
2586        tested, and the overall condition is true if the condition is true for any
2587        one of them. This is the way Perl works, and is also more like the way
2588        testing by number works.
2589    
2590    
2591    Version 7.9 11-Apr-09
2592    ---------------------
2593    
2594    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
2595        (pcretest), all targets were linked against these libraries. This included
2596        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2597        libraries. This caused unwanted dependencies to be created. This problem
2598        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2599        pcretest is linked with readline.
2600    
2601    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
2602        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2603        moved outside it again, because FALSE and TRUE are already defined in AIX,
2604        but BOOL is not.
2605    
2606    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2607        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2608    
2609    4.  The pcregrep documentation said spaces were inserted as well as colons (or
2610        hyphens) following file names and line numbers when outputting matching
2611        lines. This is not true; no spaces are inserted. I have also clarified the
2612        wording for the --colour (or --color) option.
2613    
2614    5.  In pcregrep, when --colour was used with -o, the list of matching strings
2615        was not coloured; this is different to GNU grep, so I have changed it to be
2616        the same.
2617    
2618    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
2619        each matching line was coloured. Now it goes on to look for further matches
2620        of any of the test patterns, which is the same behaviour as GNU grep.
2621    
2622    7.  A pattern that could match an empty string could cause pcregrep to loop; it
2623        doesn't make sense to accept an empty string match in pcregrep, so I have
2624        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2625        seems to be how GNU grep behaves. [But see later change 40 for release
2626        8.33.]
2627    
2628    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2629        start or after a newline", because the conditional assertion was not being
2630        correctly handled. The rule now is that both the assertion and what follows
2631        in the first alternative must satisfy the test.
2632    
2633    9.  If auto-callout was enabled in a pattern with a conditional group whose
2634        condition was an assertion, PCRE could crash during matching, both with
2635        pcre_exec() and pcre_dfa_exec().
2636    
2637    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2638        used for matching.
2639    
2640    11. Unicode property support in character classes was not working for
2641        characters (bytes) greater than 127 when not in UTF-8 mode.
2642    
2643    12. Added the -M command line option to pcretest.
2644    
2645    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2646    
2647    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2648    
2649    16. Added comments and documentation about mis-use of no_arg in the C++
2650        wrapper.
2651    
2652    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2653        from Martin Jerabek that uses macro names for all relevant character and
2654        string constants.
2655    
2656    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2657        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2658        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2659        these, but not everybody uses configure.
2660    
2661    19. A conditional group that had only one branch was not being correctly
2662        recognized as an item that could match an empty string. This meant that an
2663        enclosing group might also not be so recognized, causing infinite looping
2664        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2665        with the subject "ab", where knowledge that the repeated group can match
2666        nothing is needed in order to break the loop.
2667    
2668    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2669        exec(), but without supplying a callout function, matching went wrong.
2670    
2671    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2672        leak if the size of the offset vector was greater than 30. When the vector
2673        is smaller, the saved offsets during recursion go onto a local stack
2674        vector, but for larger vectors malloc() is used. It was failing to free
2675        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2676        error, in fact).
2677    
2678    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2679        heapframe that is used only when UTF-8 support is enabled. This caused no
2680        problem, but was untidy.
2681    
2682    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2683        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2684        included within another project.
2685    
2686    24. Steven Van Ingelgem's patches to add more options to the CMake support,
2687        slightly modified by me:
2688    
2689          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2690              not building pcregrep.
2691    
2692          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2693              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2694    
2695    25. Forward references, both numeric and by name, in patterns that made use of
2696        duplicate group numbers, could behave incorrectly or give incorrect errors,
2697        because when scanning forward to find the reference group, PCRE was not
2698        taking into account the duplicate group numbers. A pattern such as
2699        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2700    
2701    26. Changed a few more instances of "const unsigned char *" to USPTR, making
2702        the feature of a custom pointer more persuasive (as requested by a user).
2703    
2704    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2705        pcretest.c, inside #ifndefs, because it seems they are sometimes already
2706        pre-defined.
2707    
2708    28. Added support for (*UTF8) at the start of a pattern.
2709    
2710    29. Arrange for flags added by the "release type" setting in CMake to be shown
2711        in the configuration summary.
2712    
2713    
2714    Version 7.8 05-Sep-08
2715    ---------------------
2716    
2717    1.  Replaced UCP searching code with optimized version as implemented for Ad
2718        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2719        stage table and inline lookup instead of a function, giving speed ups of 2
2720        to 5 times on some simple patterns that I tested. Permission was given to
2721        distribute the MultiStage2.py script that generates the tables (it's not in
2722        the tarball, but is in the Subversion repository).
2723    
2724    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2725        scripts.
2726    
2727    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2728        a group with a zero qualifier. The result of the study could be incorrect,
2729        or the function might crash, depending on the pattern.
2730    
2731    4.  Caseless matching was not working for non-ASCII characters in back
2732        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2733        It now works when Unicode Property Support is available.
2734    
2735    5.  In pcretest, an escape such as \x{de} in the data was always generating
2736        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2737        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2738        truncation.
2739    
2740    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2741    
2742    7.  Added two (int) casts to pcregrep when printing the difference of two
2743        pointers, in case they are 64-bit values.
2744    
2745    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
2746        test 2 if it fails.
2747    
2748    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2749        and a #define of that name to empty if it is not externally set. This is to
2750        allow users of MSVC to set it if necessary.
2751    
2752    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2753        the convenience functions in the pcre_get.c source file.
2754    
2755    11. An option change at the start of a pattern that had top-level alternatives
2756        could cause overwriting and/or a crash. This command provoked a crash in
2757        some environments:
2758    
2759          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2760    
2761        This potential security problem was recorded as CVE-2008-2371.
2762    
2763    12. For a pattern where the match had to start at the beginning or immediately
2764        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2765        pcre_dfa_exec() could read past the end of the passed subject if there was
2766        no match. To help with detecting such bugs (e.g. with valgrind), I modified
2767        pcretest so that it places the subject at the end of its malloc-ed buffer.
2768    
2769    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2770        exec() might read past the end of the data buffer in UTF-8 mode.
2771    
2772    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2773        the data contained the byte 0x85 as part of a UTF-8 character within its
2774        first line. This applied both to normal and DFA matching.
2775    
2776    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2777        /^[^d]*?$/8 failed to match "abc".
2778    
2779    16. Added a missing copyright notice to pcrecpp_internal.h.
2780    
2781    17. Make it more clear in the documentation that values returned from
2782        pcre_exec() in ovector are byte offsets, not character counts.
2783    
2784    18. Tidied a few places to stop certain compilers from issuing warnings.
2785    
2786    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2787        supplied by Stefan Weber. I made a further small update for 7.8 because
2788        there is a change of source arrangements: the pcre_searchfuncs.c module is
2789        replaced by pcre_ucd.c.
2790    
2791    
2792    Version 7.7 07-May-08
2793    ---------------------
2794    
2795    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
2796        a string to a long long, pretend we don't even have a long long." This is
2797        done by checking for the strtoq, strtoll, and _strtoi64 functions.
2798    
2799    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2800        pre-7.6 versions, which defined a global no_arg variable instead of putting
2801        it in the RE class. (See also #8 below.)
2802    
2803    3.  Remove a line of dead code, identified by coverity and reported by Nuno
2804        Lopes.
2805    
2806    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
2807    
2808        (1) The include/exclude patterns were being applied to the whole pathnames
2809            of files, instead of just to the final components.
2810    
2811        (2) If there was more than one level of directory, the subdirectories were
2812            skipped unless they satisfied the include/exclude conditions. This is
2813            inconsistent with GNU grep (and could even be seen as contrary to the
2814            pcregrep specification - which I improved to make it absolutely clear).
2815            The action now is always to scan all levels of directory, and just
2816            apply the include/exclude patterns to regular files.
2817    
2818    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2819        --exclude_dir in the tests to avoid scanning .svn directories.
2820    
2821    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
2822        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2823        doesn't support NULs in patterns.
2824    
2825    7.  Added some missing "const"s to declarations of static tables in
2826        pcre_compile.c and pcre_dfa_exec.c.
2827    
2828    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2829        caused by fix #2  above. (Subsequently also a second patch to fix the
2830        first patch. And a third patch - this was a messy problem.)
2831    
2832    9.  Applied Craig's patch to remove the use of push_back().
2833    
2834    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2835        matching function regexec().
2836    
2837    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2838        which, however, unlike Perl's \g{...}, are subroutine calls, not back
2839        references. PCRE supports relative numbers with this syntax (I don't think
2840        Oniguruma does).
2841    
2842    12. Previously, a group with a zero repeat such as (...){0} was completely
2843        omitted from the compiled regex. However, this means that if the group
2844        was called as a subroutine from elsewhere in the pattern, things went wrong
2845        (an internal error was given). Such groups are now left in the compiled
2846        pattern, with a new opcode that causes them to be skipped at execution
2847        time.
2848    
2849    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2850        to the way PCRE behaves:
2851    
2852        (a) A lone ] character is dis-allowed (Perl treats it as data).
2853    
2854        (b) A back reference to an unmatched subpattern matches an empty string
2855            (Perl fails the current match path).
2856    
2857        (c) A data ] in a character class must be notated as \] because if the
2858            first data character in a class is ], it defines an empty class. (In
2859            Perl it is not possible to have an empty class.) The empty class []
2860            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2861            The negative empty class [^] matches any one character, independently
2862            of the DOTALL setting.
2863    
2864    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2865        non-existent subpattern following a character class starting with ']' and
2866        containing () gave an internal compiling error instead of "reference to
2867        non-existent subpattern". Fortunately, when the pattern did exist, the
2868        compiled code was correct. (When scanning forwards to check for the
2869        existence of the subpattern, it was treating the data ']' as terminating
2870        the class, so got the count wrong. When actually compiling, the reference
2871        was subsequently set up correctly.)
2872    
2873    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2874        it was being rejected as not supported by pcre_dfa_exec(), even though
2875        other assertions are supported. I have made pcre_dfa_exec() support
2876        (*FAIL).
2877    
2878    16. The implementation of 13c above involved the invention of a new opcode,
2879        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2880        cannot be changed at match time, I realized I could make a small
2881        improvement to matching performance by compiling OP_ALLANY instead of
2882        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2883        on the OP_ANY path.
2884    
2885    17. Compiling pcretest on Windows with readline support failed without the
2886        following two fixes: (1) Make the unistd.h include conditional on
2887        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2888    
2889    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2890        ncurses library to be included for pcretest when ReadLine support is
2891        requested, but also to allow for it to be overridden. This patch came from
2892        Daniel Bergström.
2893    
2894    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2895        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2896        any errors with the current Unicode tables. Thanks to Peter Kankowski for
2897        spotting this.
2898    
2899    
2900    Version 7.6 28-Jan-08
2901    ---------------------
2902    
2903    1.  A character class containing a very large number of characters with
2904        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2905        overflow.
2906    
2907    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
2908        HAVE_LONG_LONG is not defined.
2909    
2910    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
2911        bring them up to date and include new features. This patch includes:
2912    
2913        - Fixed PH's badly added libz and libbz2 support.
2914        - Fixed a problem with static linking.
2915        - Added pcredemo. [But later removed - see 7 below.]
2916        - Fixed dftables problem and added an option.
2917        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2918            HAVE_LONG_LONG.
2919        - Added readline support for pcretest.
2920        - Added an listing of the option settings after cmake has run.
2921    
2922    4.  A user submitted a patch to Makefile that makes it easy to create
2923        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2924        Makefile.am that cause it to include this special target, without
2925        affecting anything else. Note that the same mingw target plus all
2926        the other distribution libraries and programs are now supported
2927        when configuring with CMake (see 6 below) instead of with
2928        Configure/Make.
2929    
2930    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2931        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2932        exported in the Windows port". It has not yet been confirmed that the patch
2933        solves the problem, but it does no harm.
2934    
2935    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2936        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2937        with CMake, and also correct the comment about stack recursion.
2938    
2939    7.  Remove the automatic building of pcredemo from the ./configure system and
2940        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2941        of a program that users should build themselves after PCRE is installed, so
2942        building it automatically is not really right. What is more, it gave
2943        trouble in some build environments.
2944    
2945    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2946    
2947    
2948    Version 7.5 10-Jan-08
2949    ---------------------
2950    
2951    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2952        values in parens when parsing an RE using the C++ wrapper."
2953    
2954    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2955        Characters greater than 255 were excluded from the class instead of being
2956        included.
2957    
2958    3.  The same bug as (2) above applied to negated POSIX classes such as
2959        [:^space:].
2960    
2961    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2962        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2963        I have changed it.
2964    
2965    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2966        first named subpattern) and a construct such as (?&a) would reference the
2967        first named subpattern whose name started with "a" (in other words, the
2968        length check was missing). Both these problems are fixed. "Subpattern name
2969        expected" is now given for (?&) (a zero-length name), and this patch also
2970        makes it give the same error for \k'' (previously it complained that that
2971        was a reference to a non-existent subpattern).
2972    
2973    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2974        this is right because (?- can be followed by option settings as well as by
2975        digits. I have, however, made the messages clearer.
2976    
2977    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2978        than the number used in the conditional) now cause a compile-time error.
2979        This is actually not compatible with Perl, which accepts such patterns, but
2980        treats the conditional as always being FALSE (as PCRE used to), but it
2981        seems to me that giving a diagnostic is better.
2982    
2983    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2984        and messages.
2985    
2986    9.  Fix two occurrences of "backslash" in comments that should have been
2987        "backspace".
2988    
2989    10. Remove two redundant lines of code that can never be obeyed (their function
2990        was moved elsewhere).
2991    
2992    11. The program that makes PCRE's Unicode character property table had a bug
2993        which caused it to generate incorrect table entries for sequences of
2994        characters that have the same character type, but are in different scripts.
2995        It amalgamated them into a single range, with the script of the first of
2996        them. In other words, some characters were in the wrong script. There were
2997        thirteen such cases, affecting characters in the following ranges:
2998    
2999          U+002b0 - U+002c1
3000          U+0060c - U+0060d
3001          U+0061e - U+00612
3002          U+0064b - U+0065e
3003          U+0074d - U+0076d
3004          U+01800 - U+01805
3005          U+01d00 - U+01d77
3006          U+01d9b - U+01dbf
3007          U+0200b - U+0200f
3008          U+030fc - U+030fe
3009          U+03260 - U+0327f
3010          U+0fb46 - U+0fbb1
3011          U+10450 - U+1049d
3012    
3013    12. The -o option (show only the matching part of a line) for pcregrep was not
3014        compatible with GNU grep in that, if there was more than one match in a
3015        line, it showed only the first of them. It now behaves in the same way as
3016        GNU grep.
3017    
3018    13. If the -o and -v options were combined for pcregrep, it printed a blank
3019        line for every non-matching line. GNU grep prints nothing, and pcregrep now
3020        does the same. The return code can be used to tell if there were any
3021        non-matching lines.
3022    
3023    14. Added --file-offsets and --line-offsets to pcregrep.
3024    
3025    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
3026        infinitely looping recursion. The bug was that positive lookaheads were not
3027        being skipped when checking for a possible empty match (negative lookaheads
3028        and both kinds of lookbehind were skipped).
3029    
3030    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
3031        inclusion of <windows.h> to before rather than after the definition of
3032        INVALID_FILE_ATTRIBUTES (patch from David Byron).
3033    
3034    17. Specifying a possessive quantifier with a specific limit for a Unicode
3035        character property caused pcre_compile() to compile bad code, which led at
3036        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
3037        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
3038        caused the error; without that there was no problem.
3039    
3040    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
3041    
3042    19. Added --enable-pcretest-libreadline.
3043    
3044    20. In pcrecpp.cc, the variable 'count' was incremented twice in
3045        RE::GlobalReplace(). As a result, the number of replacements returned was
3046        double what it should be. I removed one of the increments, but Craig sent a
3047        later patch that removed the other one (the right fix) and added unit tests
3048        that check the return values (which was not done before).
3049    
3050    21. Several CMake things:
3051    
3052        (1) Arranged that, when cmake is used on Unix, the libraries end up with
3053            the names libpcre and libpcreposix, not just pcre and pcreposix.
3054    
3055        (2) The above change means that pcretest and pcregrep are now correctly
3056            linked with the newly-built libraries, not previously installed ones.
3057    
3058        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
3059    
3060    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
3061        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
3062        UTF-8 newline character). The key issue is that the pattern starts .*;
3063        this means that the match must be either at the beginning, or after a
3064        newline. The bug was in the code for advancing after a failed match and
3065        checking that the new position followed a newline. It was not taking
3066        account of UTF-8 characters correctly.
3067    
3068    23. PCRE was behaving differently from Perl in the way it recognized POSIX
3069        character classes. PCRE was not treating the sequence [:...:] as a
3070        character class unless the ... were all letters. Perl, however, seems to
3071        allow any characters between [: and :], though of course it rejects as
3072        unknown any "names" that contain non-letters, because all the known class
3073        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
3074        for example, whereas PCRE did not - it did not recognize a POSIX character
3075        class. This seemed a bit dangerous, so the code has been changed to be
3076        closer to Perl. The behaviour is not identical to Perl, because PCRE will
3077        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
3078        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
3079        Perl does, and where it didn't before.
3080    
3081    24. Rewrite so as to remove the single use of %n from pcregrep because in some
3082        Windows environments %n is disabled by default.
3083    
3084    
3085    Version 7.4 21-Sep-07
3086    ---------------------
3087    
3088    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
3089        means that a class such as [\s] counted as "explicit reference to CR or
3090        LF". That isn't really right - the whole point of the change was to try to
3091        help when there was an actual mention of one of the two characters. So now
3092        the change happens only if \r or \n (or a literal CR or LF) character is
3093        encountered.
3094    
3095    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
3096        of both had grown to the point where there were only 3 bits left.
3097        Fortunately, there was spare space in the data structure, and so I have
3098        moved the internal flags into a new 16-bit field to free up more option
3099        bits.
3100    
3101    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3102        but did not set the internal JCHANGED flag - either of these is enough to
3103        control the way the "get" function works - but the PCRE_INFO_JCHANGED
3104        facility is supposed to tell if (?J) was ever used, so now (?J) at the
3105        start sets both bits.
3106    
3107    4.  Added options (at build time, compile time, exec time) to change \R from
3108        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3109    
3110    5.  doc/pcresyntax.html was missing from the distribution.
3111    
3112    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3113        compatibility, even though it is no longer used.
3114    
3115    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3116        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3117        windows.h file is present (where different names are used). [This was
3118        reversed later after testing - see 16 below.]
3119    
3120    8.  Changed all #include <config.h> to #include "config.h". There were also
3121        some further <pcre.h> cases that I changed to "pcre.h".
3122    
3123    9.  When pcregrep was used with the --colour option, it missed the line ending
3124        sequence off the lines that it output.
3125    
3126    10. It was pointed out to me that arrays of string pointers cause lots of
3127        relocations when a shared library is dynamically loaded. A technique of
3128        using a single long string with a table of offsets can drastically reduce
3129        these. I have refactored PCRE in four places to do this. The result is
3130        dramatic:
3131    
3132          Originally:                          290
3133          After changing UCP table:            187
3134          After changing error message table:   43
3135          After changing table of "verbs"       36
3136          After changing table of Posix names   22
3137    
3138        Thanks to the folks working on Gregex for glib for this insight.
3139    
3140    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3141        unicode-properties was also set.
3142    
3143    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3144    
3145    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3146        checked only for CRLF.
3147    
3148    14. Added casts to pcretest.c to avoid compiler warnings.
3149    
3150    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3151    
3152    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3153        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3154        entirely. This removes changes made in 7 above.
3155    
3156    17. The CMake files have been updated, and there is now more information about
3157        building with CMake in the NON-UNIX-USE document.
3158    
3159    
3160    Version 7.3 28-Aug-07
3161    ---------------------
3162    
3163     1. In the rejigging of the build system that eventually resulted in 7.1, the
3164        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3165        brackets there is not right, since it causes compilers to look for an
3166        installed pcre.h, not the version that is in the source that is being
3167        compiled (which of course may be different). I have changed it back to:
3168    
3169          #include "pcre.h"
3170    
3171        I have a vague recollection that the change was concerned with compiling in
3172        different directories, but in the new build system, that is taken care of
3173        by the VPATH setting the Makefile.
3174    
3175     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3176        when the subject happened to end in the byte 0x85 (e.g. if the last
3177        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3178        characters but of course it shouldn't be taken as a newline when it is part
3179        of another character. The bug was that, for an unlimited repeat of . in
3180        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3181        characters when looking for a newline.
3182    
3183     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3184    
3185     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3186        in debug output.
3187    
3188     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3189        long printing in the pcrecpp unittest when running under MinGW.
3190    
3191     6. ESC_K was left out of the EBCDIC table.
3192    
3193     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3194        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3195        limit also applies to "virtual nesting" when a pattern is recursive, and in
3196        this case 1000 isn't so big. I have been able to remove this limit at the
3197        expense of backing off one optimization in certain circumstances. Normally,
3198        when pcre_exec() would call its internal match() function recursively and
3199        immediately return the result unconditionally, it uses a "tail recursion"
3200        feature to save stack. However, when a subpattern that can match an empty
3201        string has an unlimited repetition quantifier, it no longer makes this
3202        optimization. That gives it a stack frame in which to save the data for
3203        checking that an empty string has been matched. Previously this was taken
3204        from the 1000-entry workspace that had been reserved. So now there is no
3205        explicit limit, but more stack is used.
3206    
3207     8. Applied Daniel's patches to solve problems with the import/export magic
3208        syntax that is required for Windows, and which was going wrong for the
3209        pcreposix and pcrecpp parts of the library. These were overlooked when this
3210        problem was solved for the main library.
3211    
3212     9. There were some crude static tests to avoid integer overflow when computing
3213        the size of patterns that contain repeated groups with explicit upper
3214        limits. As the maximum quantifier is 65535, the maximum group length was
3215        set at 30,000 so that the product of these two numbers did not overflow a
3216        32-bit integer. However, it turns out that people want to use groups that
3217        are longer than 30,000 bytes (though not repeat them that many times).
3218        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3219        made it possible to implement the integer overflow checks in a much more
3220        dynamic way, which I have now done. The artificial limitation on group
3221        length has been removed - we now have only the limit on the total length of
3222        the compiled pattern, which depends on the LINK_SIZE setting.
3223    
3224    10. Fixed a bug in the documentation for get/copy named substring when
3225        duplicate names are permitted. If none of the named substrings are set, the
3226        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3227        empty string.
3228    
3229    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3230        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3231        because the ] is interpreted as the first data character and the
3232        terminating ] is not found. PCRE has been made compatible with Perl in this
3233        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3234        cause memory overwriting.
3235    
3236    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3237        string has been matched (to stop an infinite loop). It was not recognizing
3238        a conditional subpattern that could match an empty string if that
3239        subpattern was within another subpattern. For example, it looped when
3240        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
3241        condition was not nested. This bug has been fixed.
3242    
3243    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3244        past the start of the subject in the presence of bytes with the top bit
3245        set, for example "\x8aBCD".
3246    
3247    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3248        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3249    
3250    14. Optimized (?!) to (*FAIL).
3251    
3252    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3253        This restricts code points to be within the range 0 to 0x10FFFF, excluding
3254        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3255        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3256        does: it's just the validity check that is more restrictive.
3257    
3258    16. Inserted checks for integer overflows during escape sequence (backslash)
3259        processing, and also fixed erroneous offset values for syntax errors during
3260        backslash processing.
3261    
3262    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3263        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3264    
3265    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3266        caused an overrun.
3267    
3268    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3269        something other than just ASCII characters) inside a group that had an
3270        unlimited repeat caused a loop at compile time (while checking to see
3271        whether the group could match an empty string).
3272    
3273    20. Debugging a pattern containing \p or \P could cause a crash. For example,
3274        [\P{Any}] did so. (Error in the code for printing property names.)
3275    
3276    21. An orphan \E inside a character class could cause a crash.
3277    
3278    22. A repeated capturing bracket such as (A)? could cause a wild memory
3279        reference during compilation.
3280    
3281    23. There are several functions in pcre_compile() that scan along a compiled
3282        expression for various reasons (e.g. to see if it's fixed length for look
3283        behind). There were bugs in these functions when a repeated \p or \P was
3284        present in the pattern. These operators have additional parameters compared
3285        with \d, etc, and these were not being taken into account when moving along
3286        the compiled data. Specifically:
3287    
3288        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3289            length.
3290    
3291        (b) An item such as \pL+ within a repeated group could cause crashes or
3292            loops.
3293    
3294        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3295            "reference to non-existent subpattern" error.
3296    
3297        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3298    
3299    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3300        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3301    
3302    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3303    
3304    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3305        character were causing crashes (broken optimization).
3306    
3307    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3308        \p or \P) caused a compile-time loop.
3309    
3310    28. More problems have arisen in unanchored patterns when CRLF is a valid line
3311        break. For example, the unstudied pattern [\r\n]A does not match the string
3312        "\r\nA" because change 7.0/46 below moves the current point on by two
3313        characters after failing to match at the start. However, the pattern \nA
3314        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3315        the same is true. There doesn't seem any very clean way out of this, but
3316        what I have chosen to do makes the common cases work: PCRE now takes note
3317        of whether there can be an explicit match for \r or \n anywhere in the
3318        pattern, and if so, 7.0/46 no longer applies. As part of this change,
3319        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3320        pattern has explicit CR or LF references.
3321    
3322    29. Added (*CR) etc for changing newline setting at start of pattern.
3323    
3324    
3325    Version 7.2 19-Jun-07
3326  ---------------------  ---------------------
3327    
3328   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,   1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
# Line 21  Version 7.2 01-May-07 Line 3342  Version 7.2 01-May-07
3342      stack recursion. This gives a massive performance boost under BSD, but just      stack recursion. This gives a massive performance boost under BSD, but just
3343      a small improvement under Linux. However, it saves one field in the frame      a small improvement under Linux. However, it saves one field in the frame
3344      in all cases.      in all cases.
3345    
3346   6. Added more features from the forthcoming Perl 5.10:   6. Added more features from the forthcoming Perl 5.10:
3347    
3348      (a) (?-n) (where n is a string of digits) is a relative subroutine or      (a) (?-n) (where n is a string of digits) is a relative subroutine or
3349          recursion call. It refers to the nth most recently opened parentheses.          recursion call. It refers to the nth most recently opened parentheses.
3350    
3351      (b) (?+n) is also a relative subroutine call; it refers to the nth next      (b) (?+n) is also a relative subroutine call; it refers to the nth next
3352          to be opened parentheses.          to be opened parentheses.
3353    
3354      (c) Conditions that refer to capturing parentheses can be specified      (c) Conditions that refer to capturing parentheses can be specified
3355          relatively, for example, (?(-2)... or (?(+3)...          relatively, for example, (?(-2)... or (?(+3)...
3356    
3357        (d) \K resets the start of the current match so that everything before
3358            is not part of it.
3359    
3360        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3361    
3362        (f) \g{name} is another synonym - part of Perl 5.10's unification of
3363            reference syntax.
3364    
3365        (g) (?| introduces a group in which the numbering of parentheses in each
3366            alternative starts with the same number.
3367    
3368        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3369    
3370     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3371        PCRE_INFO_JCHANGED.
3372    
3373     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
3374        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3375        for detecting groups that can match an empty string.
3376    
3377