ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1680 - (show annotations)
Tue Feb 14 08:48:18 2017 UTC (3 years ago) by zherczeg
File size: 292273 byte(s)
Error occurred while calculating annotation data.
Fix a missing else in the JIT compiler reported by 'idaifish'.
1 ChangeLog for PCRE
2 ------------------
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
7 Version 8.41
8 ------------
10 1. Fixed typo in CMakeLists.txt (wrong number of arguments for
11 PCRE_STATIC_RUNTIME (affects MSVC only).
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.
18 3. Fix a missing else in the JIT compiler reported by 'idaifish'.
21 Version 8.40 11-January-2017
22 ----------------------------
24 1. Using -o with -M in pcregrep could cause unnecessary repeated output when
25 the match extended over a line boundary.
27 2. Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
28 MSVC static compilation, putting the first patch under a new option.
30 3. Fix register overwite in JIT when SSE2 acceleration is enabled.
32 4. Ignore "show all captures" (/=) for DFA matching.
34 5. Fix JIT unaligned accesses on x86. Patch by Marc Mutz.
36 6. In any wide-character mode (8-bit UTF or any 16-bit or 32-bit mode),
37 without PCRE_UCP set, a negative character type such as \D in a positive
38 class should cause all characters greater than 255 to match, whatever else
39 is in the class. There was a bug that caused this not to happen if a
40 Unicode property item was added to such a class, for example [\D\P{Nd}] or
41 [\W\pL].
43 7. When pcretest was outputing information from a callout, the caret indicator
44 for the current position in the subject line was incorrect if it was after
45 an escape sequence for a character whose code point was greater than
46 \x{ff}.
48 8. A pattern such as (?<RA>abc)(?(R)xyz) was incorrectly compiled such that
49 the conditional was interpreted as a reference to capturing group 1 instead
50 of a test for recursion. Any group whose name began with R was
51 misinterpreted in this way. (The reference interpretation should only
52 happen if the group's name is precisely "R".)
54 9. A number of bugs have been mended relating to match start-up optimizations
55 when the first thing in a pattern is a positive lookahead. These all
56 applied only when PCRE_NO_START_OPTIMIZE was *not* set:
58 (a) A pattern such as (?=.*X)X$ was incorrectly optimized as if it needed
59 both an initial 'X' and a following 'X'.
60 (b) Some patterns starting with an assertion that started with .* were
61 incorrectly optimized as having to match at the start of the subject or
62 after a newline. There are cases where this is not true, for example,
63 (?=.*[A-Z])(?=.{8,16})(?!.*[\s]) matches after the start in lines that
64 start with spaces. Starting .* in an assertion is no longer taken as an
65 indication of matching at the start (or after a newline).
68 Version 8.39 14-June-2016
69 -------------------------
71 1. If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
72 an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
73 misbehaved. This bug was found by the LLVM fuzzer.
75 2. Similar to the above, if an isolated \E was present between an item and its
76 qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
77 bug was found by the LLVM fuzzer.
79 3. Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
80 working correctly in UCP mode.
82 4. The POSIX wrapper function regexec() crashed if the option REG_STARTEND
83 was set when the pmatch argument was NULL. It now returns REG_INVARG.
85 5. Allow for up to 32-bit numbers in the ordin() function in pcregrep.
87 6. An empty \Q\E sequence between an item and its qualifier caused
88 pcre_compile() to misbehave when auto callouts were enabled. This bug was
89 found by the LLVM fuzzer.
91 7. If a pattern that was compiled with PCRE_EXTENDED started with white
92 space or a #-type comment that was followed by (?-x), which turns off
93 PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
94 pcre_compile() assumed that (?-x) applied to the whole pattern and
95 consequently mis-compiled it. This bug was found by the LLVM fuzzer.
97 8. A call of pcre_copy_named_substring() for a named substring whose number
98 was greater than the space in the ovector could cause a crash.
100 9. Yet another buffer overflow bug involved duplicate named groups with a
101 group that reset capture numbers (compare 8.38/7 below). Once again, I have
102 just allowed for more memory, even if not needed. (A proper fix is
103 implemented in PCRE2, but it involves a lot of refactoring.)
105 10. pcre_get_substring_list() crashed if the use of \K in a match caused the
106 start of the match to be earlier than the end.
108 11. Migrating appropriate PCRE2 JIT improvements to PCRE.
110 12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
111 assertion, caused pcretest to generate incorrect output, and also to read
112 uninitialized memory (detected by ASAN or valgrind).
114 13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
115 nested set of parentheses of sufficient size caused an overflow of the
116 compiling workspace (which was diagnosed, but of course is not desirable).
118 14. And yet another buffer overflow bug involving duplicate named groups, this
119 time nested, with a nested back reference. Yet again, I have just allowed
120 for more memory, because anything more needs all the refactoring that has
121 been done for PCRE2. An example pattern that provoked this bug is:
122 /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
123 registered as CVE-2016-1283.
125 15. pcretest went into a loop if global matching was requested with an ovector
126 size less than 2. It now gives an error message. This bug was found by
127 afl-fuzz.
129 16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
130 ("assertion expected") when (?(?C) was not followed by an opening
131 parenthesis.
133 17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
134 actually affect anything, by sheer luck.
136 18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
137 static compilation.
139 19. Modified the RunTest script to incorporate a valgrind suppressions file so
140 that certain errors, provoked by the SSE2 instruction set when JIT is used,
141 are ignored.
143 20. A racing condition is fixed in JIT reported by Mozilla.
145 21. Minor code refactor to avoid "array subscript is below array bounds"
146 compiler warning.
148 22. Minor code refactor to avoid "left shift of negative number" warning.
150 23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
151 UCP support.
153 24. Refactor to avoid compiler warnings in pcrecpp.cc.
155 25. Refactor to fix a typo in pcre_jit_test.c
157 26. Patch to support compiling pcrecpp.cc with Intel compiler.
160 Version 8.38 23-November-2015
161 -----------------------------
163 1. If a group that contained a recursive back reference also contained a
164 forward reference subroutine call followed by a non-forward-reference
165 subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
166 compile correct code, leading to undefined behaviour or an internally
167 detected error. This bug was discovered by the LLVM fuzzer.
169 2. Quantification of certain items (e.g. atomic back references) could cause
170 incorrect code to be compiled when recursive forward references were
171 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
172 This bug was discovered by the LLVM fuzzer.
174 3. A repeated conditional group whose condition was a reference by name caused
175 a buffer overflow if there was more than one group with the given name.
176 This bug was discovered by the LLVM fuzzer.
178 4. A recursive back reference by name within a group that had the same name as
179 another group caused a buffer overflow. For example:
180 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
182 5. A forward reference by name to a group whose number is the same as the
183 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
184 a buffer overflow at compile time. This bug was discovered by the LLVM
185 fuzzer.
187 6. A lookbehind assertion within a set of mutually recursive subpatterns could
188 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
190 7. Another buffer overflow bug involved duplicate named groups with a
191 reference between their definition, with a group that reset capture
192 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
193 fixed by always allowing for more memory, even if not needed. (A proper fix
194 is implemented in PCRE2, but it involves more refactoring.)
196 8. There was no check for integer overflow in subroutine calls such as (?123).
198 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
199 being treated as a literal 'l' instead of causing an error.
201 10. There was a buffer overflow if pcre_exec() was called with an ovector of
202 size 1. This bug was found by american fuzzy lop.
204 11. If a non-capturing group containing a conditional group that could match
205 an empty string was repeated, it was not identified as matching an empty
206 string itself. For example: /^(?:(?(1)x|)+)+$()/.
208 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
209 \a and \e in test subject lines.
211 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
212 instead of the EBCDIC value.
214 14. The handling of \c in an EBCDIC environment has been revised so that it is
215 now compatible with the specification in Perl's perlebcdic page.
217 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
218 ASCII/Unicode. This has now been added to the list of characters that are
219 recognized as white space in EBCDIC.
221 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
222 error (correctly) when used outside a class, but did not give an error
223 within a class.
225 17. \h within a class was incorrectly compiled in EBCDIC environments.
227 18. A pattern with an unmatched closing parenthesis that contained a backward
228 assertion which itself contained a forward reference caused buffer
229 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
231 19. JIT should return with error when the compiled pattern requires more stack
232 space than the maximum.
234 20. A possessively repeated conditional group that could match an empty string,
235 for example, /(?(R))*+/, was incorrectly compiled.
237 21. Fix infinite recursion in the JIT compiler when certain patterns such as
238 /(?:|a|){100}x/ are analysed.
240 22. Some patterns with character classes involving [: and \\ were incorrectly
241 compiled and could cause reading from uninitialized memory or an incorrect
242 error diagnosis.
244 23. Pathological patterns containing many nested occurrences of [: caused
245 pcre_compile() to run for a very long time.
247 24. A conditional group with only one branch has an implicit empty alternative
248 branch and must therefore be treated as potentially matching an empty
249 string.
251 25. If (?R was followed by - or + incorrect behaviour happened instead of a
252 diagnostic.
254 26. Arrange to give up on finding the minimum matching length for overly
255 complex patterns.
257 27. Similar to (4) above: in a pattern with duplicated named groups and an
258 occurrence of (?| it is possible for an apparently non-recursive back
259 reference to become recursive if a later named group with the relevant
260 number is encountered. This could lead to a buffer overflow. Wen Guanxing
261 from Venustech ADLAB discovered this bug.
263 28. If pcregrep was given the -q option with -c or -l, or when handling a
264 binary file, it incorrectly wrote output to stdout.
266 29. The JIT compiler did not restore the control verb head in case of *THEN
267 control verbs. This issue was found by Karl Skomski with a custom LLVM
268 fuzzer.
270 30. Error messages for syntax errors following \g and \k were giving inaccurate
271 offsets in the pattern.
273 31. Added a check for integer overflow in conditions (?(<digits>) and
274 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
275 fuzzer.
277 32. Handling recursive references such as (?2) when the reference is to a group
278 later in the pattern uses code that is very hacked about and error-prone.
279 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
280 give an internal error if it is obvious that compiling has gone wrong.
282 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
283 This issue was found by Karl Skomski with a custom LLVM fuzzer.
285 34. The JIT compiler should restore the control chain for empty possessive
286 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
288 35. Match limit check added to JIT recursion. This issue was found by Karl
289 Skomski with a custom LLVM fuzzer.
291 36. Yet another case similar to 27 above has been circumvented by an
292 unconditional allocation of extra memory. This issue is fixed "properly" in
293 PCRE2 by refactoring the way references are handled. Wen Guanxing
294 from Venustech ADLAB discovered this bug.
296 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
297 with a custom LLVM fuzzer.
299 38. Fixed a corner case of range optimization in JIT.
301 39. An incorrect error "overran compiling workspace" was given if there were
302 exactly enough group forward references such that the last one extended
303 into the workspace safety margin. The next one would have expanded the
304 workspace. The test for overflow was not including the safety margin.
306 40. A match limit issue is fixed in JIT which was found by Karl Skomski
307 with a custom LLVM fuzzer.
309 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
310 work under Windows. (Why has it taken so long for anyone to notice?)
312 42. In a character class such as [\W\p{Any}] where both a negative-type escape
313 ("not a word character") and a property escape were present, the property
314 escape was being ignored.
316 43. Fix crash caused by very long (*MARK) or (*THEN) names.
318 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
319 by a single ASCII character in a class item, was incorrectly compiled in
320 UCP mode. The POSIX class got lost, but only if the single character
321 followed it.
323 45. [:punct:] in UCP mode was matching some characters in the range 128-255
324 that should not have been matched.
326 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
327 class, all characters with code points greater than 255 are in the class.
328 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
329 such as \w are turned into Unicode properties), wide characters were not
330 correctly handled, and could fail to match.
333 Version 8.37 28-April-2015
334 --------------------------
336 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
337 for those parentheses to be closed with whatever has been captured so far.
338 However, it was failing to mark any other groups between the hightest
339 capture so far and the currrent group as "unset". Thus, the ovector for
340 those groups contained whatever was previously there. An example is the
341 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
343 2. If an assertion condition was quantified with a minimum of zero (an odd
344 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
346 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
347 unrecognized modifier, a crash could occur.
349 4. An attempt to do global matching in pcretest with a zero-length ovector
350 caused a crash.
352 5. Fixed a memory leak during matching that could occur for a subpattern
353 subroutine call (recursive or otherwise) if the number of captured groups
354 that had to be saved was greater than ten.
356 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
357 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
358 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
360 7. A UTF pattern containing a "not" match of a non-ASCII character and a
361 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
363 8. When a pattern is compiled, it remembers the highest back reference so that
364 when matching, if the ovector is too small, extra memory can be obtained to
365 use instead. A conditional subpattern whose condition is a check on a
366 capture having happened, such as, for example in the pattern
367 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
368 setting the highest backreference number. This mattered only if pcre_exec()
369 was called with an ovector that was too small to hold the capture, and there
370 was no other kind of back reference (a situation which is probably quite
371 rare). The effect of the bug was that the condition was always treated as
372 FALSE when the capture could not be consulted, leading to a incorrect
373 behaviour by pcre_exec(). This bug has been fixed.
375 9. A reference to a duplicated named group (either a back reference or a test
376 for being set in a conditional) that occurred in a part of the pattern where
377 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
378 to be incorrectly calculated, leading to overwriting.
380 10. A mutually recursive set of back references such as (\2)(\1) caused a
381 segfault at study time (while trying to find the minimum matching length).
382 The infinite loop is now broken (with the minimum length unset, that is,
383 zero).
385 11. If an assertion that was used as a condition was quantified with a minimum
386 of zero, matching went wrong. In particular, if the whole group had
387 unlimited repetition and could match an empty string, a segfault was
388 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
389 assertions to be quantified, but not if they are being used as conditions,
390 so the above pattern is faulted by Perl. PCRE has now been changed so that
391 it also rejects such patterns.
393 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
394 failed to allow the zero-repeat case if pcre2_exec() was called with an
395 ovector too small to capture the group.
397 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
398 Red Hat Product Security:
400 (a) A crash if /K and /F were both set with the option to save the compiled
401 pattern.
403 (b) Another crash if the option to print captured substrings in a callout
404 was combined with setting a null ovector, for example \O\C+ as a subject
405 string.
407 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
408 forward reference repeated a large (but limited) number of times within a
409 repeated outer group that has a zero minimum quantifier, caused incorrect
410 code to be compiled, leading to the error "internal error:
411 previously-checked referenced subpattern not found" when an incorrect
412 memory address was read. This bug was reported as "heap overflow",
413 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
414 CVE-2015-2325.
416 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
417 call within a group that also contained a recursive back reference caused
418 incorrect code to be compiled. This bug was reported as "heap overflow",
419 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
420 number CVE-2015-2326.
422 24. Computing the size of the JIT read-only data in advance has been a source
423 of various issues, and new ones are still appear unfortunately. To fix
424 existing and future issues, size computation is eliminated from the code,
425 and replaced by on-demand memory allocation.
427 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
428 adjacent to the end of the range, and the range contained characters with
429 more than one other case, caused incorrect behaviour when compiled in UTF
430 mode. In that example, the range a-j was left out of the class.
432 26. Fix JIT compilation of conditional blocks, which assertion
433 is converted to (*FAIL). E.g: /(?(?!))/.
435 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
436 discovered by the LLVM fuzzer.
438 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
439 when this assertion was used as a condition, for example (?(?!)a|b). In
440 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
441 error about an unsupported item.
443 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
444 possessification code could take exponential time to complete. A recursion
445 depth limit of 1000 has been imposed to limit the resources used by this
446 optimization.
448 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
449 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
450 because \S ensures they are all in the class. The code for doing this was
451 interacting badly with the code for computing the amount of space needed to
452 compile the pattern, leading to a buffer overflow. This bug was discovered
453 by the LLVM fuzzer.
455 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
456 other kinds of group caused stack overflow at compile time. This bug was
457 discovered by the LLVM fuzzer.
459 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
460 between a subroutine call and its quantifier was incorrectly compiled,
461 leading to buffer overflow or other errors. This bug was discovered by the
462 LLVM fuzzer.
464 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
465 assertion after (?(. The code was failing to check the character after
466 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
467 was discovered by the LLVM fuzzer.
469 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
470 a fixed maximum following a group that contains a subroutine reference was
471 incorrectly compiled and could trigger buffer overflow. This bug was
472 discovered by the LLVM fuzzer.
474 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
475 caused a stack overflow instead of the diagnosis of a non-fixed length
476 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
478 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
479 (e.g. /(?<=\Ka)/) could make pcregrep loop.
481 37. There was a similar problem to 36 in pcretest for global matches.
483 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
484 and a subsequent item in the pattern caused a non-match, backtracking over
485 the repeated \X did not stop, but carried on past the start of the subject,
486 causing reference to random memory and/or a segfault. There were also some
487 other cases where backtracking after \C could crash. This set of bugs was
488 discovered by the LLVM fuzzer.
490 39. The function for finding the minimum length of a matching string could take
491 a very long time if mutual recursion was present many times in a pattern,
492 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
493 method has been implemented. This infelicity was discovered by the LLVM
494 fuzzer.
496 40. Static linking against the PCRE library using the pkg-config module was
497 failing on missing pthread symbols.
500 Version 8.36 26-September-2014
501 ------------------------------
503 1. Got rid of some compiler warnings in the C++ modules that were shown up by
504 -Wmissing-field-initializers and -Wunused-parameter.
506 2. The tests for quantifiers being too big (greater than 65535) were being
507 applied after reading the number, and stupidly assuming that integer
508 overflow would give a negative number. The tests are now applied as the
509 numbers are read.
511 3. Tidy code in pcre_exec.c where two branches that used to be different are
512 now the same.
514 4. The JIT compiler did not generate match limit checks for certain
515 bracketed expressions with quantifiers. This may lead to exponential
516 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
517 issue should be resolved now.
519 5. Fixed an issue, which occures when nested alternatives are optimized
520 with table jumps.
522 6. Inserted two casts and changed some ints to size_t in the light of some
523 reported 64-bit compiler warnings (Bugzilla 1477).
525 7. Fixed a bug concerned with zero-minimum possessive groups that could match
526 an empty string, which sometimes were behaving incorrectly in the
527 interpreter (though correctly in the JIT matcher). This pcretest input is
528 an example:
530 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
533 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
534 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
535 for an empty string was breaking the inner loop and carrying on at a lower
536 level, when possessive repeated groups should always return to a higher
537 level as they have no backtrack points in them. The empty string test now
538 occurs at the outer level.
540 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
541 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
543 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
544 doesn't).
546 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
547 to hold a pattern to be incorrectly computed (too small) when there were
548 named back references to duplicated names. This could cause "internal
549 error: code overflow" or "double free or corruption" or other memory
550 handling errors.
552 11. When named subpatterns had the same prefixes, back references could be
553 confused. For example, in this pattern:
555 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
557 the reference to 'Name' was incorrectly treated as a reference to a
558 duplicate name.
560 12. A pattern such as /^s?c/mi8 where the optional character has more than
561 one "other case" was incorrectly compiled such that it would only try to
562 match starting at "c".
564 13. When a pattern starting with \s was studied, VT was not included in the
565 list of possible starting characters; this should have been part of the
566 8.34/18 patch.
568 14. If a character class started [\Qx]... where x is any character, the class
569 was incorrectly terminated at the ].
571 15. If a pattern that started with a caseless match for a character with more
572 than one "other case" was studied, PCRE did not set up the starting code
573 unit bit map for the list of possible characters. Now it does. This is an
574 optimization improvement, not a bug fix.
576 16. The Unicode data tables have been updated to Unicode 7.0.0.
578 17. Fixed a number of memory leaks in pcregrep.
580 18. Avoid a compiler warning (from some compilers) for a function call with
581 a cast that removes "const" from an lvalue by using an intermediate
582 variable (to which the compiler does not object).
584 19. Incorrect code was compiled if a group that contained an internal recursive
585 back reference was optional (had quantifier with a minimum of zero). This
586 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
587 caused segmentation faults because of stack overflows at compile time.
589 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
590 group that is quantified with an indefinite repeat, caused a compile-time
591 loop which used up all the system stack and provoked a segmentation fault.
592 This was not the same bug as 19 above.
594 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
595 Patch by Mike Frysinger.
598 Version 8.35 04-April-2014
599 --------------------------
601 1. A new flag is set, when property checks are present in an XCLASS.
602 When this flag is not set, PCRE can perform certain optimizations
603 such as studying these XCLASS-es.
605 2. The auto-possessification of character sets were improved: a normal
606 and an extended character set can be compared now. Furthermore
607 the JIT compiler optimizes more character set checks.
609 3. Got rid of some compiler warnings for potentially uninitialized variables
610 that show up only when compiled with -O2.
612 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
613 of a match later then the end of the match. The pcretest program was not
614 handling the case sensibly - it was outputting from the start to the next
615 binary zero. It now reports this situation in a message, and outputs the
616 text from the end to the start.
618 5. Fast forward search is improved in JIT. Instead of the first three
619 characters, any three characters with fixed position can be searched.
620 Search order: first, last, middle.
622 6. Improve character range checks in JIT. Characters are read by an inprecise
623 function now, which returns with an unknown value if the character code is
624 above a certain threshold (e.g: 256). The only limitation is that the value
625 must be bigger than the threshold as well. This function is useful when
626 the characters above the threshold are handled in the same way.
628 7. The macros whose names start with RAWUCHAR are placeholders for a future
629 mode in which only the bottom 21 bits of 32-bit data items are used. To
630 make this more memorable for those maintaining the code, the names have
631 been changed to start with UCHAR21, and an extensive comment has been added
632 to their definition.
634 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
635 to the export list in Makefile.am (they were accidentally omitted from the
636 8.34 tarball).
638 9. The informational output from pcretest used the phrase "starting byte set"
639 which is inappropriate for the 16-bit and 32-bit libraries. As the output
640 for "first char" and "need char" really means "non-UTF-char", I've changed
641 "byte" to "char", and slightly reworded the output. The documentation about
642 these values has also been (I hope) clarified.
644 10. Another JIT related optimization: use table jumps for selecting the correct
645 backtracking path, when more than four alternatives are present inside a
646 bracket.
648 11. Empty match is not possible, when the minimum length is greater than zero,
649 and there is no \K in the pattern. JIT should avoid empty match checks in
650 such cases.
652 12. In a caseless character class with UCP support, when a character with more
653 than one alternative case was not the first character of a range, not all
654 the alternative cases were added to the class. For example, s and \x{17f}
655 are both alternative cases for S: the class [RST] was handled correctly,
656 but [R-T] was not.
658 13. The configure.ac file always checked for pthread support when JIT was
659 enabled. This is not used in Windows, so I have put this test inside a
660 check for the presence of windows.h (which was already tested for).
662 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
663 The algorithm provides a way to skip certain starting offsets, and usually
664 faster than linear prefix searches.
666 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
667 as for 'fr_FR' and 'french'. For some reason, however, it then used the
668 Windows-specific input and output files, which have 'french' screwed in.
669 So this could never have worked. One of the problems with locales is that
670 they aren't always the same. I have now updated RunTest so that it checks
671 the output of the locale test (test 3) against three different output
672 files, and it allows the test to pass if any one of them matches. With luck
673 this should make the test pass on some versions of Solaris where it was
674 failing. Because of the uncertainty, the script did not used to stop if
675 test 3 failed; it now does. If further versions of a French locale ever
676 come to light, they can now easily be added.
678 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
679 there was a message during ./configure, but it did not stop. This now
680 provokes an error. The invalid example in README has been corrected.
681 If a value less than the minimum is given, the minimum value has always
682 been used, but now a warning is given.
684 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
685 was a bug in the test system, which is now fixed. Also, the list of various
686 configurations that are tested for each release did not have one with both
687 16/32 bits and --enable-bar-anycrlf. It now does.
689 18. pcretest was missing "-C bsr" for displaying the \R default setting.
691 19. Little endian PowerPC systems are supported now by the JIT compiler.
693 20. The fast forward newline mechanism could enter to an infinite loop on
694 certain invalid UTF-8 input. Although we don't support these cases
695 this issue can be fixed by a performance optimization.
697 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
698 not take account if existing stack usage. There is now a new global
699 variable called pcre_stack_guard that can be set to point to an external
700 function to check stack availability. It is called at the start of
701 processing every parenthesized group.
703 22. A typo in the code meant that in ungreedy mode the max/min qualifier
704 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
705 not match "ab".
707 23. When UTF was disabled, the JIT program reported some incorrect compile
708 errors. These messages are silenced now.
710 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
711 compiler.
713 25. Change all the temporary files used in RunGrepTest to be different to those
714 used by RunTest so that the tests can be run simultaneously, for example by
715 "make -j check".
718 Version 8.34 15-December-2013
719 -----------------------------
721 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
722 executable memory. Patch inspired by Carsten Klein.
724 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
725 this macro is never tested and has no effect, because the work to support
726 coverage involves only compiling and linking options and special targets in
727 the Makefile. The comment in config.h implied that defining the macro would
728 enable coverage support, which is totally false. There was also support for
729 setting this macro in the CMake files (my fault, I just copied it from
730 configure). SUPPORT_GCOV has now been removed.
732 3. Make a small performance improvement in strlen16() and strlen32() in
733 pcretest.
735 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
736 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
737 these cases). There was also one in pcretest.c.
739 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
741 6. In UTF mode, the code for checking whether a group could match an empty
742 string (which is used for indefinitely repeated groups to allow for
743 breaking an infinite loop) was broken when the group contained a repeated
744 negated single-character class with a character that occupied more than one
745 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
746 UTF-8 mode). The effect was undefined: the group might or might not be
747 deemed as matching an empty string, or the program might have crashed.
749 7. The code for checking whether a group could match an empty string was not
750 recognizing that \h, \H, \v, \V, and \R must match a character.
752 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
753 an empty string. If it can, pcretest shows this in its information output.
755 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
756 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
757 matched by pcre_exec() without using JIT:
759 (a) If the rest of the pattern did not match after a maximal run of
760 grapheme clusters, the code for backing up to try with fewer of them
761 did not always back up over a full grapheme when characters that do not
762 have the modifier quality were involved, e.g. Hangul syllables.
764 (b) If the match point in a subject started with modifier character, and
765 there was no match, the code could incorrectly back up beyond the match
766 point, and potentially beyond the first character in the subject,
767 leading to a segfault or an incorrect match result.
769 10. A conditional group with an assertion condition could lead to PCRE
770 recording an incorrect first data item for a match if no other first data
771 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
772 first data item, and therefore matched "ca" after "c" instead of at the
773 start.
775 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
776 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
778 12. The source of pcregrep now includes z/OS-specific code so that it can be
779 compiled for z/OS as part of the special z/OS distribution.
781 13. Added the -T and -TM options to pcretest.
783 14. The code in pcre_compile.c for creating the table of named capturing groups
784 has been refactored. Instead of creating the table dynamically during the
785 actual compiling pass, the information is remembered during the pre-compile
786 pass (on the stack unless there are more than 20 named groups, in which
787 case malloc() is used) and the whole table is created before the actual
788 compile happens. This has simplified the code (it is now nearly 150 lines
789 shorter) and prepared the way for better handling of references to groups
790 with duplicate names.
792 15. A back reference to a named subpattern when there is more than one of the
793 same name now checks them in the order in which they appear in the pattern.
794 The first one that is set is used for the reference. Previously only the
795 first one was inspected. This change makes PCRE more compatible with Perl.
797 16. Unicode character properties were updated from Unicode 6.3.0.
799 17. The compile-time code for auto-possessification has been refactored, based
800 on a patch by Zoltan Herczeg. It now happens after instead of during
801 compilation. The code is cleaner, and more cases are handled. The option
802 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
803 options in pcretest are provided to set it. It can also be set by
804 (*NO_AUTO_POSSESS) at the start of a pattern.
806 18. The character VT has been added to the default ("C" locale) set of
807 characters that match \s and are generally treated as white space,
808 following this same change in Perl 5.18. There is now no difference between
809 "Perl space" and "POSIX space". Whether VT is treated as white space in
810 other locales depends on the locale.
812 19. The code for checking named groups as conditions, either for being set or
813 for being recursed, has been refactored (this is related to 14 and 15
814 above). Processing unduplicated named groups should now be as fast at
815 numerical groups, and processing duplicated groups should be faster than
816 before.
818 20. Two patches to the CMake build system, by Alexander Barkov:
820 (1) Replace the "source" command by "." in CMakeLists.txt because
821 "source" is a bash-ism.
823 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
824 without these the CMake build does not work on Solaris.
826 21. Perl has changed its handling of \8 and \9. If there is no previously
827 encountered capturing group of those numbers, they are treated as the
828 literal characters 8 and 9 instead of a binary zero followed by the
829 literals. PCRE now does the same.
831 22. Following Perl, added \o{} to specify codepoints in octal, making it
832 possible to specify values greater than 0777 and also making them
833 unambiguous.
835 23. Perl now gives an error for missing closing braces after \x{... instead of
836 treating the string as literal. PCRE now does the same.
838 24. RunTest used to grumble if an inappropriate test was selected explicitly,
839 but just skip it when running all tests. This make it awkward to run ranges
840 of tests when one of them was inappropriate. Now it just skips any
841 inappropriate tests, as it always did when running all tests.
843 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
844 character types such as \d or \w, too many callouts were inserted, and the
845 data that they returned was rubbish.
847 26. In UCP mode, \s was not matching two of the characters that Perl matches,
848 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
849 were matched by \h. The code has now been refactored so that the lists of
850 the horizontal and vertical whitespace characters used for \h and \v (which
851 are defined only in one place) are now also used for \s.
853 27. Add JIT support for the 64 bit TileGX architecture.
854 Patch by Jiong Wang (Tilera Corporation).
856 28. Possessive quantifiers for classes (both explicit and automatically
857 generated) now use special opcodes instead of wrapping in ONCE brackets.
859 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
860 (because it's meaningless), this was not happening when PCRE_CASELESS was
861 set. Not wrong, but inefficient.
863 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
864 properties for \w, \d, etc) is present in a test regex. Otherwise if the
865 test contains no characters greater than 255, Perl doesn't realise it
866 should be using Unicode semantics.
868 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
869 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
870 does in Unicode mode.
872 32. Added the "forbid" facility to pcretest so that putting tests into the
873 wrong test files can sometimes be quickly detected.
875 33. There is now a limit (default 250) on the depth of nesting of parentheses.
876 This limit is imposed to control the amount of system stack used at compile
877 time. It can be changed at build time by --with-parens-nest-limit=xxx or
878 the equivalent in CMake.
880 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
881 errors. Perl warns for these when in warning mode, but PCRE has no facility
882 for giving warnings.
884 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
885 However, this was not working for (?!) because it is optimized to (*FAIL),
886 for which PCRE does not allow quantifiers. The optimization is now disabled
887 when a quantifier follows (?!). I can't see any use for this, but it makes
888 things uniform.
890 36. Perl no longer allows group names to start with digits, so I have made this
891 change also in PCRE. It simplifies the code a bit.
893 37. In extended mode, Perl ignores spaces before a + that indicates a
894 possessive quantifier. PCRE allowed a space before the quantifier, but not
895 before the possessive +. It now does.
897 38. The use of \K (reset reported match start) within a repeated possessive
898 group such as (a\Kb)*+ was not working.
900 40. Document that the same character tables must be used at compile time and
901 run time, and that the facility to pass tables to pcre_exec() and
902 pcre_dfa_exec() is for use only with saved/restored patterns.
904 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
905 features for Builds with MSVC:
907 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
908 builds). The .rc files can be used to set FileDescription and many other
909 attributes.
911 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
912 This allows higher-level build scripts which want .pdb files to avoid
913 hard-coding the exact files needed."
915 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
916 mean "start of word" and "end of word", respectively, as a transition aid.
918 43. A minimizing repeat of a class containing codepoints greater than 255 in
919 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
920 compiled to use the heap for recursion.
922 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
923 is configured.
926 Version 8.33 28-May-2013
927 ------------------------
929 1. Added 'U' to some constants that are compared to unsigned integers, to
930 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
931 variables that are added to signed variables, to ensure the result is
932 signed and can be negated.
934 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
935 CMake config files.
937 3. Revise the creation of config.h.generic so that all boolean macros are
938 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
939 overriding via -D on the command line possible.
941 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
942 to unsigned int is reported to make a quite noticeable speed difference in
943 a specific Windows environment. Testing on Linux did also appear to show
944 some benefit (and it is clearly not harmful). Also fixed the definition of
945 Xop which should be unsigned.
947 5. Related to (4), changing the definition of the intermediate variable cc
948 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
949 improvements.
951 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
952 unnecessary spaces.
954 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
955 and later.
957 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
959 9. Optimizing fast_forward_start_bits in JIT.
961 10. Adding support for callouts in JIT, and fixing some issues revealed
962 during this work. Namely:
964 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
966 (b) Minimum length was not checked before the matching is started.
968 11. The value of capture_last that is passed to callouts was incorrect in some
969 cases when there was a capture on one path that was subsequently abandoned
970 after a backtrack. Also, the capture_last value is now reset after a
971 recursion, since all captures are also reset in this case.
973 12. The interpreter no longer returns the "too many substrings" error in the
974 case when an overflowing capture is in a branch that is subsequently
975 abandoned after a backtrack.
977 13. In the pathological case when an offset vector of size 2 is used, pcretest
978 now prints out the matched string after a yield of 0 or 1.
980 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
981 Only supported by the JIT compiler at the moment.
983 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
985 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
986 offset of the starting point of the matching process, provided the offsets
987 vector is large enough.
989 17. The \A escape now records a lookbehind value of 1, though its execution
990 does not actually inspect the previous character. This is to ensure that,
991 in partial multi-segment matching, at least one character from the old
992 segment is retained when a new segment is processed. Otherwise, if there
993 are no lookbehinds in the pattern, \A might match incorrectly at the start
994 of a new segment.
996 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
998 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
999 gives some modest performance improvement in 8-bit mode.
1001 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
1002 be expressed in certain programming languages using Universal Character
1003 Names.
1005 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
1006 which points out that "non characters" are not "characters that may not
1007 appear in Unicode strings" but rather "characters that are reserved for
1008 internal use and have only local meaning".
1010 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
1011 there was a conditional group that depended on an assertion, if the
1012 assertion was false, the callout that immediately followed the alternation
1013 in the condition was skipped when pcre_exec() was used for matching.
1015 23. Allow an explicit callout to be inserted before an assertion that is the
1016 condition for a conditional group, for compatibility with automatic
1017 callouts, which always insert a callout at this point.
1019 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
1020 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
1022 25. (*PRUNE) is now supported by the JIT compiler.
1024 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
1026 27. Fix the case where there are two or more SKIPs with arguments that may be
1027 ignored.
1029 28. (*SKIP) is now supported by the JIT compiler.
1031 29. (*THEN) is now supported by the JIT compiler.
1033 30. Update RunTest with additional test selector options.
1035 31. The way PCRE handles backtracking verbs has been changed in two ways.
1037 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
1038 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
1039 backtracking. In some cases this makes it more Perl-compatible, but Perl's
1040 rather obscure rules do not always do the same thing.
1042 (2) Previously, backtracking verbs were confined within assertions. This is
1043 no longer the case for positive assertions, except for (*ACCEPT). Again,
1044 this sometimes improves Perl compatibility, and sometimes does not.
1046 32. A number of tests that were in test 2 because Perl did things differently
1047 have been moved to test 1, because either Perl or PCRE has changed, and
1048 these tests are now compatible.
1050 32. Backtracking control verbs are now handled in the same way in JIT and
1051 interpreter.
1053 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1054 contained a forward subroutine reference caused a compile error.
1056 34. Auto-detect and optimize limited repetitions in JIT.
1058 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1059 blocking (*UTF) etc.
1061 36. In the interpreter, maximizing pattern repetitions for characters and
1062 character types now use tail recursion, which reduces stack usage.
1064 37. The value of the max lookbehind was not correctly preserved if a compiled
1065 and saved regex was reloaded on a host of different endianness.
1067 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1068 of the compiled pattern block, expand the flags field from 16 to 32 bits
1069 because it was almost full.
1071 39. Try madvise first before posix_madvise.
1073 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1074 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1075 original change locked out all matches of empty strings. This has been
1076 changed so that one match of an empty string per line is recognized.
1077 Subsequent searches on the same line (for colouring or for --only-matching,
1078 for example) do not recognize empty strings.
1080 41. Applied a user patch to fix a number of spelling mistakes in comments.
1082 42. Data lines longer than 65536 caused pcretest to crash.
1084 43. Clarified the data type for length and startoffset arguments for pcre_exec
1085 and pcre_dfa_exec in the function-specific man pages, where they were
1086 explicitly stated to be in bytes, never having been updated. I also added
1087 some clarification to the pcreapi man page.
1089 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1090 a segmentation fault.
1093 Version 8.32 30-November-2012
1094 -----------------------------
1096 1. Improved JIT compiler optimizations for first character search and single
1097 character iterators.
1099 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1100 Patch by Daniel Richard G.
1102 3. Single character iterator optimizations in the JIT compiler.
1104 4. Improved JIT compiler optimizations for character ranges.
1106 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1107 Reported by Giuseppe D'Angelo.
1109 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1110 of a line, was being set incorrectly in cases where .* appeared inside
1111 atomic brackets at the start of a pattern, or where there was a subsequent
1112 *PRUNE or *SKIP.
1114 7. Improved instruction cache flush for POWER/PowerPC.
1115 Patch by Daniel Richard G.
1117 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
1118 grep:
1120 (a) There is now no limit to the number of patterns to be matched.
1122 (b) An error is given if a pattern is too long.
1124 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1125 are now supported.
1127 (d) --exclude-from and --include-from (multiple use) have been added.
1129 (e) Exclusions and inclusions now apply to all files and directories, not
1130 just to those obtained from scanning a directory recursively.
1132 (f) Multiple uses of -f and --file-list are now supported.
1134 (g) In a Windows environment, the default for -d has been changed from
1135 "read" (the GNU grep default) to "skip", because otherwise the presence
1136 of a directory in the file list provokes an error.
1138 (h) The documentation has been revised and clarified in places.
1140 9. Improve the matching speed of capturing brackets.
1142 10. Changed the meaning of \X so that it now matches a Unicode extended
1143 grapheme cluster.
1145 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1146 out POSIX threads when JIT support is configured.
1148 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1150 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1151 structure could go wrong in environments where size_t is not the same size
1152 as int.
1154 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1155 set.
1157 15. The EBCDIC support had decayed; later updates to the code had included
1158 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1159 general tidy up of EBCDIC-related issues, and the documentation was also
1160 not quite right. There is now a test that can be run on ASCII systems to
1161 check some of the EBCDIC-related things (but is it not a full test).
1163 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1164 in a small tidy to the code.
1166 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1168 18. If the --only-matching (-o) option in pcregrep is specified multiple
1169 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1170 substrings matched by the 1st and 2nd capturing parentheses. A separating
1171 string can be specified by --om-separator (default empty).
1173 19. Improving the first n character searches.
1175 20. Turn case lists for horizontal and vertical white space into macros so that
1176 they are defined only once.
1178 21. This set of changes together give more compatible Unicode case-folding
1179 behaviour for characters that have more than one other case when UCP
1180 support is available.
1182 (a) The Unicode property table now has offsets into a new table of sets of
1183 three or more characters that are case-equivalent. The MultiStage2.py
1184 script that generates these tables (the pcre_ucd.c file) now scans
1185 CaseFolding.txt instead of UnicodeData.txt for character case
1186 information.
1188 (b) The code for adding characters or ranges of characters to a character
1189 class has been abstracted into a generalized function that also handles
1190 case-independence. In UTF-mode with UCP support, this uses the new data
1191 to handle characters with more than one other case.
1193 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1194 whose other case is greater than 256 are now correctly matched
1195 caselessly. Previously, the high codepoint matched the low one, but not
1196 vice versa.
1198 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1199 of the new class addition function, using character lists defined as
1200 macros alongside the case definitions of 20 above.
1202 (e) Caseless back references now work with characters that have more than
1203 one other case.
1205 (f) General caseless matching of characters with more than one other case
1206 is supported.
1208 22. Unicode character properties were updated from Unicode 6.2.0
1210 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1212 24. Add support for 32-bit character strings, and UTF-32
1214 25. Major JIT compiler update (code refactoring and bugfixing).
1215 Experimental Sparc 32 support is added.
1217 26. Applied a modified version of Daniel Richard G's patch to create
1218 pcre.h.generic and config.h.generic by "make" instead of in the
1219 PrepareRelease script.
1221 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1222 pcre_compile.c when checking for a zero character.
1224 28. Introducing a native interface for JIT. Through this interface, the compiled
1225 machine code can be directly executed. The purpose of this interface is to
1226 provide fast pattern matching, so several sanity checks are not performed.
1227 However, feature tests are still performed. The new interface provides
1228 1.4x speedup compared to the old one.
1230 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1231 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1232 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1234 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1235 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1236 the "old" RFC 2279). Instead, it ended up passing a negative length to
1237 pcre_exec().
1239 31. Add support for GCC's visibility feature to hide internal functions.
1241 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1242 "unknown -C option" after outputting 0 or 1.
1244 33. There is now support for generating a code coverage report for the test
1245 suite in environments where gcc is the compiler and lcov is installed. This
1246 is mainly for the benefit of the developers.
1248 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1249 unaddressable using valgrind annotations, allowing valgrind to detect
1250 invalid memory accesses. This is mainly for the benefit of the developers.
1252 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1254 26. Give configure error if --enable-cpp but no C++ compiler found.
1257 Version 8.31 06-July-2012
1258 -------------------------
1260 1. Fixing a wrong JIT test case and some compiler warnings.
1262 2. Removed a bashism from the RunTest script.
1264 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1265 to unsigned type, result still unsigned" that was given by an MS compiler
1266 on encountering the code "-sizeof(xxx)".
1268 4. Partial matching support is added to the JIT compiler.
1270 5. Fixed several bugs concerned with partial matching of items that consist
1271 of more than one character:
1273 (a) /^(..)\1/ did not partially match "aba" because checking references was
1274 done on an "all or nothing" basis. This also applied to repeated
1275 references.
1277 (b) \R did not give a hard partial match if \r was found at the end of the
1278 subject.
1280 (c) \X did not give a hard partial match after matching one or more
1281 characters at the end of the subject.
1283 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1284 a partial match for the string "\r".
1286 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1287 a partial match for a CR character at the end of the subject string.
1289 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1290 running pcretest, the text "(JIT)" added to the output whenever JIT is
1291 actually used to run the match.
1293 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1294 or /S+[+] with a digit between 1 and 7.
1296 8. OP_NOT now supports any UTF character not just single-byte ones.
1298 9. (*MARK) control verb is now supported by the JIT compiler.
1300 10. The command "./RunTest list" lists the available tests without actually
1301 running any of them. (Because I keep forgetting what they all are.)
1305 12. Applied a (slightly modified) user-supplied patch that improves performance
1306 when the heap is used for recursion (compiled with --disable-stack-for-
1307 recursion). Instead of malloc and free for each heap frame each time a
1308 logical recursion happens, frames are retained on a chain and re-used where
1309 possible. This sometimes gives as much as 30% improvement.
1311 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1312 call.
1314 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1316 15. It is now possible to link pcretest with libedit as an alternative to
1317 libreadline.
1319 16. (*COMMIT) control verb is now supported by the JIT compiler.
1321 17. The Unicode data tables have been updated to Unicode 6.1.0.
1323 18. Added --file-list option to pcregrep.
1325 19. Added binary file support to pcregrep, including the -a, --binary-files,
1326 -I, and --text options.
1328 20. The madvise function is renamed for posix_madvise for QNX compatibility
1329 reasons. Fixed by Giuseppe D'Angelo.
1331 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1333 22. Changed the option for creating symbolic links for 16-bit man pages from
1334 -s to -sf so that re-installing does not cause issues.
1336 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1338 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1339 matches in certain environments (the workspace was not being correctly
1340 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1341 some of the workspace data at the beginning of a restart.
1343 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1344 was not doing so when it should - probably a typo introduced by SVN 528
1345 (change 8.10/14).
1347 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1348 \w+ when the character tables indicated that \x{c4} was a word character.
1349 There were several related cases, all because the tests for doing a table
1350 lookup were testing for characters less than 127 instead of 255.
1352 27. If a pattern contains capturing parentheses that are not used in a match,
1353 their slots in the ovector are set to -1. For those that are higher than
1354 any matched groups, this happens at the end of processing. In the case when
1355 there were back references that the ovector was too small to contain
1356 (causing temporary malloc'd memory to be used during matching), and the
1357 highest capturing number was not used, memory off the end of the ovector
1358 was incorrectly being set to -1. (It was using the size of the temporary
1359 memory instead of the true size.)
1361 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1362 ovector size, it uses memory at the end of the block that it has got.
1364 29. Check for an overlong MARK name and give an error at compile time. The
1365 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1367 30. JIT compiler update.
1369 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1370 Rill for the patch.
1372 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1374 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1376 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1377 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1378 was enabled.
1380 35. Improve JIT code generation for greedy plus quantifier.
1382 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1383 1 to "aa" instead of to an empty string. The bug affected repeated groups
1384 that could potentially match an empty string.
1386 37. Optimizing single character iterators in JIT.
1388 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1389 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1390 codepoints that are too big for the mode are faulted, and in a UTF mode,
1391 disallowed codepoints are also faulted.
1393 39. If PCRE was compiled with UTF support, in three places in the DFA
1394 matcher there was code that should only have been obeyed in UTF mode, but
1395 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1396 processing when bytes with values greater than 127 were present. In 16-bit
1397 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1398 both cases the values are those that cannot be the first data item in a UTF
1399 character. The three items that might have provoked this were recursions,
1400 possessively repeated groups, and atomic groups.
1402 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1403 and pcregrep, because some OS require shared objects to be explicitly
1404 passed to ld, causing the link step to fail if they are not.
1406 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1407 mode, patterns that started with \h* or \R* might be incorrectly matched.
1410 Version 8.30 04-February-2012
1411 -----------------------------
1413 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1414 name is defined in ctype.h.
1416 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1417 only in quite long subpatterns.
1419 3. Removed the function pcre_info(), which has been obsolete and deprecated
1420 since it was replaced by pcre_fullinfo() in February 2000.
1422 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1423 match a (*MARK), and the match failed at the start of the subject, a
1424 reference to memory before the start of the subject could occur. This bug
1425 was introduced by fix 17 of release 8.21.
1427 5. A reference to an unset group with zero minimum repetition was giving
1428 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1429 /(another)?(\1?)test/ matched against "hello world test". This bug was
1430 introduced in release 8.13.
1432 6. Add support for 16-bit character strings (a large amount of work involving
1433 many changes and refactorings).
1435 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1436 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1437 from a file.
1439 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1440 rounding is not applied in this particular case).
1442 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1443 if they appear, or are escaped, in patterns.
1445 10. Get rid of a number of -Wunused-but-set-variable warnings.
1447 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1448 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1449 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1450 also returns the mark "x". This bug applied to capturing parentheses,
1451 non-capturing parentheses, and atomic parentheses. It also applied to some
1452 assertions.
1454 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1455 information out of configure.ac instead of relying on pcre.h.generic, which
1456 is not stored in the repository.
1458 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1459 -lreadline.
1461 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1463 15. Applied Graycode's patch to put the top-level frame on the stack rather
1464 than the heap when not using the stack for recursion. This gives a
1465 performance improvement in many cases when recursion is not deep.
1467 16. Experimental code added to "pcretest -C" to output the stack frame size.
1470 Version 8.21 12-Dec-2011
1471 ------------------------
1473 1. Updating the JIT compiler.
1475 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1476 are added as well.
1478 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1479 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1480 calling _pcre_jit_exec. Some extra comments are added.
1482 4. (*MARK) settings inside atomic groups that do not contain any capturing
1483 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1484 was introduced by change 18 for 8.20.
1486 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1487 ECMA-262 standard.
1489 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1490 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1491 This bug was probably introduced by change 9 of 8.13.
1493 7. While fixing 6 above, I noticed that a number of other items were being
1494 incorrectly rejected as "not fixed length". This arose partly because newer
1495 opcodes had not been added to the fixed-length checking code. I have (a)
1496 corrected the bug and added tests for these items, and (b) arranged for an
1497 error to occur if an unknown opcode is encountered while checking for fixed
1498 length instead of just assuming "not fixed length". The items that were
1499 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1500 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1501 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1503 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1504 being incorrectly compiled and would have given unpredicatble results.
1506 9. A possessively repeated subpattern with minimum repeat count greater than
1507 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1508 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1509 the first (A) could occur when it should not.
1511 10. Add a cast and remove a redundant test from the code.
1513 11. JIT should use pcre_malloc/pcre_free for allocation.
1515 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1516 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1517 is anything other than /usr, -L is still shown).
1519 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1521 14. Perl does not support \N without a following name in a [] class; PCRE now
1522 also gives an error.
1524 15. If a forward reference was repeated with an upper limit of around 2000,
1525 it caused the error "internal error: overran compiling workspace". The
1526 maximum number of forward references (including repeats) was limited by the
1527 internal workspace, and dependent on the LINK_SIZE. The code has been
1528 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1529 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1530 of around 200,000 forward references. While doing this, I also speeded up
1531 the filling in of repeated forward references.
1533 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1534 incorrectly expecting the subject to contain another "a" after the start.
1536 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1537 in the match, the SKIP should be ignored. This was not happening; instead
1538 the SKIP was being treated as NOMATCH. For patterns such as
1539 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1540 tested.
1542 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1543 now much more compatible with Perl, in particular in cases where the result
1544 is a non-match for a non-anchored pattern. For example, if
1545 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1546 "m", where previously it did not return a name. A side effect of this
1547 change is that for partial matches, the last encountered mark name is
1548 returned, as for non matches. A number of tests that were previously not
1549 Perl-compatible have been moved into the Perl-compatible test files. The
1550 refactoring has had the pleasing side effect of removing one argument from
1551 the match() function, thus reducing its stack requirements.
1553 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1554 subsequent uses of /S (without +) incorrectly behaved like /S+.
1556 21. Retrieve executable code size support for the JIT compiler and fixing
1557 some warnings.
1559 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1560 not work when the shorter character appeared right at the end of the
1561 subject string.
1563 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1564 systems.
1566 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1567 output it when the /M option is used in pcretest.
1569 25. The CheckMan script was not being included in the distribution. Also, added
1570 an explicit "perl" to run Perl scripts from the PrepareRelease script
1571 because this is reportedly needed in Windows.
1573 26. If study data was being save in a file and studying had not found a set of
1574 "starts with" bytes for the pattern, the data written to the file (though
1575 never used) was taken from uninitialized memory and so caused valgrind to
1576 complain.
1578 27. Updated RunTest.bat as provided by Sheri Pierce.
1580 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1582 29. Computation of memory usage for the table of capturing group names was
1583 giving an unnecessarily large value.
1586 Version 8.20 21-Oct-2011
1587 ------------------------
1589 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1590 a POSIX class. After further experiments with Perl, which convinced me that
1591 Perl has bugs and confusions, a closing square bracket is no longer allowed
1592 in a POSIX name. This bug also affected patterns with classes that started
1593 with full stops.
1595 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1596 captured substring, but while checking the failing first alternative,
1597 substring 1 is temporarily captured. If the output vector supplied to
1598 pcre_exec() was not big enough for this capture, the yield of the function
1599 was still zero ("insufficient space for captured substrings"). This cannot
1600 be totally fixed without adding another stack variable, which seems a lot
1601 of expense for a edge case. However, I have improved the situation in cases
1602 such as /(a)(b)x|abc/ matched against "abc", where the return code
1603 indicates that fewer than the maximum number of slots in the ovector have
1604 been set.
1606 3. Related to (2) above: when there are more back references in a pattern than
1607 slots in the output vector, pcre_exec() uses temporary memory during
1608 matching, and copies in the captures as far as possible afterwards. It was
1609 using the entire output vector, but this conflicts with the specification
1610 that only 2/3 is used for passing back captured substrings. Now it uses
1611 only the first 2/3, for compatibility. This is, of course, another edge
1612 case.
1614 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1615 main code base, and can be used by building with --enable-jit. When this is
1616 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1617 runtime --no-jit option is given.
1619 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1620 ovector, the return from the function was zero, implying that there were
1621 other matches that did not fit. The correct "exactly full" value is now
1622 returned.
1624 6. If a subpattern that was called recursively or as a subroutine contained
1625 (*PRUNE) or any other control that caused it to give a non-standard return,
1626 invalid errors such as "Error -26 (nested recursion at the same subject
1627 position)" or even infinite loops could occur.
1629 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1630 computing the minimum length on reaching *ACCEPT, and so ended up with the
1631 wrong value of 1 rather than 0. Further investigation indicates that
1632 computing a minimum subject length in the presence of *ACCEPT is difficult
1633 (think back references, subroutine calls), and so I have changed the code
1634 so that no minimum is registered for a pattern that contains *ACCEPT.
1636 8. If (*THEN) was present in the first (true) branch of a conditional group,
1637 it was not handled as intended. [But see 16 below.]
1639 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1640 Sheri Pierce.
1642 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1643 the first byte in a match must be "a".
1645 11. Change 17 for 8.13 increased the recursion depth for patterns like
1646 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1647 pattern contains any instances of (*THEN). If it does not, the old
1648 optimizations are restored. It would be nice to do this on a per-group
1649 basis, but at the moment that is not feasible.
1651 12. In some environments, the output of pcretest -C is CRLF terminated. This
1652 broke RunTest's code that checks for the link size. A single white space
1653 character after the value is now allowed for.
1655 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1656 For "fr", it uses the Windows-specific input and output files.
1658 14. If (*THEN) appeared in a group that was called recursively or as a
1659 subroutine, it did not work as intended. [But see next item.]
1661 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1662 pattern fragments (but not containing any | characters). If A and B are
1663 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1664 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1665 D. In other words, Perl considers parentheses that do not contain any |
1666 characters to be part of a surrounding alternative, whereas PCRE was
1667 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1668 differently. PCRE now behaves in the same way as Perl, except in the case
1669 of subroutine/recursion calls such as (?1) which have in any case always
1670 been different (but PCRE had them first :-).
1672 16. Related to 15 above: Perl does not treat the | in a conditional group as
1673 creating alternatives. Such a group is treated in the same way as an
1674 ordinary group without any | characters when processing (*THEN). PCRE has
1675 been changed to match Perl's behaviour.
1677 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1678 RunGrepTest script failed.
1680 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1681 inevitable for groups that contain captures, but it can lead to a lot of
1682 stack use in large patterns. The old behaviour has been restored for atomic
1683 groups that do not contain any capturing parentheses.
1685 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1686 suppress the check for a minimum subject length at run time. (If it was
1687 given to pcre_exec() or pcre_dfa_exec() it did work.)
1689 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1690 fail to work when decoding hex characters in data strings in EBCDIC
1691 environments.
1693 21. It appears that in at least one Mac OS environment, the isxdigit() function
1694 is implemented as a macro that evaluates to its argument more than once,
1695 contravening the C 90 Standard (I haven't checked a later standard). There
1696 was an instance in pcretest which caused it to go wrong when processing
1697 \x{...} escapes in subject strings. The has been rewritten to avoid using
1698 things like p++ in the argument of isxdigit().
1701 Version 8.13 16-Aug-2011
1702 ------------------------
1704 1. The Unicode data tables have been updated to Unicode 6.0.0.
1706 2. Two minor typos in pcre_internal.h have been fixed.
1708 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1709 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1710 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1712 4. There were a number of related bugs in the code for matching backrefences
1713 caselessly in UTF-8 mode when codes for the characters concerned were
1714 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1715 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1716 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1717 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1718 2-byte code at the end of the subject (it thought there wasn't enough data
1719 left).
1721 5. Comprehensive information about what went wrong is now returned by
1722 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1723 as the output vector has at least 2 elements. The offset of the start of
1724 the failing character and a reason code are placed in the vector.
1726 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1727 now returned is for the first byte of the failing character, instead of the
1728 last byte inspected. This is an incompatible change, but I hope it is small
1729 enough not to be a problem. It makes the returned offset consistent with
1730 pcre_exec() and pcre_dfa_exec().
1732 7. pcretest now gives a text phrase as well as the error number when
1733 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1734 failure, the offset and reason code are output.
1736 8. When \R was used with a maximizing quantifier it failed to skip backwards
1737 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1738 back over a single character (\n). This seems wrong (because it treated the
1739 two characters as a single entity when going forwards), conflicts with the
1740 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1741 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1742 has been changed.
1744 9. Some internal refactoring has changed the processing so that the handling
1745 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1746 time (the PCRE_DOTALL option was changed this way some time ago: version
1747 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1748 which was always a bit of a fudge. It also means that there is one less
1749 argument for the match() function, which reduces its stack requirements
1750 slightly. This change also fixes an incompatibility with Perl: the pattern
1751 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1753 10. More internal refactoring has drastically reduced the number of recursive
1754 calls to match() for possessively repeated groups such as (abc)++ when
1755 using pcre_exec().
1757 11. While implementing 10, a number of bugs in the handling of groups were
1758 discovered and fixed:
1760 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1761 (a|)*(?1) gave a compile-time internal error.
1762 ((a|)+)+ did not notice that the outer group could match an empty string.
1763 (^a|^)+ was not marked as anchored.
1764 (.*a|.*)+ was not marked as matching at start or after a newline.
1766 12. Yet more internal refactoring has removed another argument from the match()
1767 function. Special calls to this function are now indicated by setting a
1768 value in a variable in the "match data" data block.
1770 13. Be more explicit in pcre_study() instead of relying on "default" for
1771 opcodes that mean there is no starting character; this means that when new
1772 ones are added and accidentally left out of pcre_study(), testing should
1773 pick them up.
1775 14. The -s option of pcretest has been documented for ages as being an old
1776 synonym of -m (show memory usage). I have changed it to mean "force study
1777 for every regex", that is, assume /S for every regex. This is similar to -i
1778 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1779 using it. It makes it easier to run collections of tests with and without
1780 study enabled, and thereby test pcre_study() more easily. All the standard
1781 tests are now run with and without -s (but some patterns can be marked as
1782 "never study" - see 20 below).
1784 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1785 restoration of the capturing data to the outer values was not happening
1786 correctly.
1788 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1789 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1790 pattern had matched an empty string, and so incorrectly returned a no
1791 match.
1793 17. There was optimizing code for the last branch of non-capturing parentheses,
1794 and also for the obeyed branch of a conditional subexpression, which used
1795 tail recursion to cut down on stack usage. Unfortunately, now that there is
1796 the possibility of (*THEN) occurring in these branches, tail recursion is
1797 no longer possible because the return has to be checked for (*THEN). These
1798 two optimizations have therefore been removed. [But see 8.20/11 above.]
1800 18. If a pattern containing \R was studied, it was assumed that \R always
1801 matched two bytes, thus causing the minimum subject length to be
1802 incorrectly computed because \R can also match just one byte.
1804 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1805 was incorrectly computed.
1807 20. If /S is present twice on a test pattern in pcretest input, it now
1808 *disables* studying, thereby overriding the use of -s on the command line
1809 (see 14 above). This is necessary for one or two tests to keep the output
1810 identical in both cases.
1812 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1813 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1815 22. When an atomic group that contained a capturing parenthesis was
1816 successfully matched, but the branch in which it appeared failed, the
1817 capturing was not being forgotten if a higher numbered group was later
1818 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1819 group 1 to "a", when in fact it should be unset. This applied to multi-
1820 branched capturing and non-capturing groups, repeated or not, and also to
1821 positive assertions (capturing in negative assertions does not happen
1822 in PCRE) and also to nested atomic groups.
1824 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1825 subject after a captured substring, to make it easier to tell which of a
1826 number of identical substrings has been captured.
1828 24. The way atomic groups are processed by pcre_exec() has been changed so that
1829 if they are repeated, backtracking one repetition now resets captured
1830 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1831 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1832 "aaa". Previously, it would have been "a". As part of this code
1833 refactoring, the way recursive calls are handled has also been changed.
1835 25. If an assertion condition captured any substrings, they were not passed
1836 back unless some other capturing happened later. For example, if
1837 (?(?=(a))a) was matched against "a", no capturing was returned.
1839 26. When studying a pattern that contained subroutine calls or assertions,
1840 the code for finding the minimum length of a possible match was handling
1841 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1842 group 1 called group 2 while simultaneously a separate group 2 called group
1843 1). A stack overflow occurred in this case. I have fixed this by limiting
1844 the recursion depth to 10.
1846 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1847 Fortmann. This supports explicit test numbers on the command line, and has
1848 argument validation and error reporting.
1850 28. An instance of \X with an unlimited repeat could fail if at any point the
1851 first character it looked at was a mark character.
1853 29. Some minor code refactoring concerning Unicode properties and scripts
1854 should reduce the stack requirement of match() slightly.
1856 30. Added the '=' option to pcretest to check the setting of unused capturing
1857 slots at the end of the pattern, which are documented as being -1, but are
1858 not included in the return count.
1860 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1861 compiled something random. Now it gives a compile-time error (as does
1862 Perl).
1864 32. A *MARK encountered during the processing of a positive assertion is now
1865 recorded and passed back (compatible with Perl).
1867 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1868 had alternative anchored branches, the search for a second match in a line
1869 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1870 matched the line "0102" twice. The same bug affected patterns that started
1871 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1872 twice.
1874 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1875 does, and because of capturing effects, quantifying parenthesized
1876 assertions may at times be useful. Quantifiers are now allowed for
1877 parenthesized assertions.
1879 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1881 36. \g was being checked for fancy things in a character class, when it should
1882 just be a literal "g".
1884 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1885 appearance of a nested POSIX class supersedes an apparent external class.
1886 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1887 unescaped square brackets may also appear as part of class names. For
1888 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1889 more like Perl. (But see 8.20/1 above.)
1891 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1892 was because it thought it was \N{name}, which is not supported).
1894 39. Add minix to OS list not supporting the -S option in pcretest.
1896 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1897 cannot analyze patterns in sufficient detail to catch mutual recursions
1898 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1899 subgroup is called recursively as a subpattern for a second time at the
1900 same position in the subject string. In previous releases this might have
1901 been caught by the recursion limit, or it might have run out of stack.
1903 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1904 happen only once. PCRE was, however incorrectly giving a compile time error
1905 "recursive call could loop indefinitely" because it cannot analyze the
1906 pattern in sufficient detail. The compile time test no longer happens when
1907 PCRE is compiling a conditional subpattern, but actual runaway loops are
1908 now caught at runtime (see 40 above).
1910 42. It seems that Perl allows any characters other than a closing parenthesis
1911 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1912 has been changed to be the same.
1914 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1915 as not to get warnings when autogen.sh is called. Also changed
1916 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1918 44. To help people who use pcregrep to scan files containing exceedingly long
1919 lines, the following changes have been made:
1921 (a) The default value of the buffer size parameter has been increased from
1922 8K to 20K. (The actual buffer used is three times this size.)
1924 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1925 PCRE is built.
1927 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1928 to be set at run time.
1930 (d) Numerical values in pcregrep options can be followed by K or M, for
1931 example --buffer-size=50K.
1933 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1934 given and the return code is set to 2.
1936 45. Add a pointer to the latest mark to the callout data block.
1938 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1939 partial match of an empty string instead of no match. This was specific to
1940 the use of ".".
1942 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1943 complete match instead of a partial match. This bug was dependent on both
1944 the PCRE_UTF8 and PCRE_DOTALL options being set.
1946 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1947 starting byte set, because \b was not being ignored.
1950 Version 8.12 15-Jan-2011
1951 ------------------------
1953 1. Fixed some typos in the markup of the man pages, and wrote a script that
1954 checks for such things as part of the documentation building process.
1956 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1957 --match-limit and --recursion-limit options (added for 8.11). In
1958 particular, this made one of the standard tests fail. (The integer value
1959 went into the wrong half of a long int.)
1961 3. If the --colour option was given to pcregrep with -v (invert match), it
1962 did strange things, either producing crazy output, or crashing. It should,
1963 of course, ignore a request for colour when reporting lines that do not
1964 match.
1966 4. Another pcregrep bug caused similar problems if --colour was specified with
1967 -M (multiline) and the pattern match finished with a line ending.
1969 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1970 matched in multiline mode, the following line was shown as part of the
1971 match. This seems wrong, so I have changed it.
1973 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1974 the check for further matches in the same line (so they could be coloured)
1975 to overrun the end of the current line. If another match was found, it was
1976 incorrectly shown (and then shown again when found in the next line).
1978 7. If pcregrep was compiled under Windows, there was a reference to the
1979 function pcregrep_exit() before it was defined. I am assuming this was
1980 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1981 reported by a user. I've moved the definition above the reference.
1984 Version 8.11 10-Dec-2010
1985 ------------------------
1987 1. (*THEN) was not working properly if there were untried alternatives prior
1988 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1989 backtracked to try for "b" instead of moving to the next alternative branch
1990 at the same level (in this case, to look for "c"). The Perl documentation
1991 is clear that when (*THEN) is backtracked onto, it goes to the "next
1992 alternative in the innermost enclosing group".
1994 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1995 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1996 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1997 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1998 (*THEN).
2000 3. If \s appeared in a character class, it removed the VT character from
2001 the class, even if it had been included by some previous item, for example
2002 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
2003 of \s, but is part of the POSIX "space" class.)
2005 4. A partial match never returns an empty string (because you can always
2006 match an empty string at the end of the subject); however the checking for
2007 an empty string was starting at the "start of match" point. This has been
2008 changed to the "earliest inspected character" point, because the returned
2009 data for a partial match starts at this character. This means that, for
2010 example, /(?<=abc)def/ gives a partial match for the subject "abc"
2011 (previously it gave "no match").
2013 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
2014 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
2015 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
2016 has an implication that the given string is incomplete (because a partial
2017 match is preferred over a full match). For this reason, these items now
2018 give a partial match in this situation. [Aside: previously, the one case
2019 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
2020 match rather than a full match, which was wrong by the old rules, but is
2021 now correct.]
2023 6. There was a bug in the handling of #-introduced comments, recognized when
2024 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
2025 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
2026 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
2027 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
2028 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
2029 places in pcre_compile().
2031 7. Related to (6) above, when pcre_compile() was skipping #-introduced
2032 comments when looking ahead for named forward references to subpatterns,
2033 the only newline sequence it recognized was NL. It now handles newlines
2034 according to the set newline convention.
2036 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
2037 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
2038 cater for a lack of strerror(). These oversights have been fixed.
2040 9. Added --match-limit and --recursion-limit to pcregrep.
2042 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
2044 11. When the -o option was used, pcregrep was setting a return code of 1, even
2045 when matches were found, and --line-buffered was not being honoured.
2047 12. Added an optional parentheses number to the -o and --only-matching options
2048 of pcregrep.
2050 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
2051 can match an empty string. The code to do it in pcretest and pcredemo
2052 needed fixing:
2054 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
2055 only one byte after an empty string match just before CRLF (this case
2056 just got forgotten; "any" and "anycrlf" were OK).
2058 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
2059 mode when an empty string match preceded an ASCII character followed by
2060 a non-ASCII character. (The code for advancing by one character rather
2061 than one byte was nonsense.)
2063 (c) The pcredemo.c sample program did not have any code at all to handle
2064 the cases when CRLF is a valid newline sequence.
2066 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2067 as a starting offset was within the subject string. There is now a new
2068 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2069 negative or greater than the length of the string. In order to test this,
2070 pcretest is extended to allow the setting of negative starting offsets.
2072 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2073 starting offset points to the beginning of a UTF-8 character was
2074 unnecessarily clumsy. I tidied it up.
2076 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2077 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2079 17. Nobody had reported that the --include_dir option, which was added in
2080 release 7.7 should have been called --include-dir (hyphen, not underscore)
2081 for compatibility with GNU grep. I have changed it to --include-dir, but
2082 left --include_dir as an undocumented synonym, and the same for
2083 --exclude-dir, though that is not available in GNU grep, at least as of
2084 release 2.5.4.
2086 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2087 characters from a string of bytes) have been redefined so as not to use
2088 loops, in order to improve performance in some environments. At the same
2089 time, I abstracted some of the common code into auxiliary macros to save
2090 repetition (this should not affect the compiled code).
2092 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2093 compile-time error is now given if \c is not followed by an ASCII
2094 character, that is, a byte less than 128. (In EBCDIC mode, the code is
2095 different, and any byte value is allowed.)
2097 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2098 START_OPTIMIZE option, which is now allowed at compile time - but just
2099 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2100 to pcregrep and other applications that have no direct access to PCRE
2101 options. The new /Y option in pcretest sets this option when calling
2102 pcre_compile().
2104 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2105 back references. Groups containing recursive back references were forced to
2106 be atomic by that change, but in the case of named groups, the amount of
2107 memory required was incorrectly computed, leading to "Failed: internal
2108 error: code overflow". This has been fixed.
2110 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2111 pcretest.c, to avoid build problems in some Borland environments.
2114 Version 8.10 25-Jun-2010
2115 ------------------------
2117 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2118 THEN.
2120 2. (*ACCEPT) was not working when inside an atomic group.
2122 3. Inside a character class, \B is treated as a literal by default, but
2123 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2124 causes the error). The code is unchanged, but I tidied the documentation.
2126 4. Inside a character class, PCRE always treated \R and \X as literals,
2127 whereas Perl faults them if its -w option is set. I have changed PCRE so
2128 that it faults them when PCRE_EXTRA is set.
2130 5. Added support for \N, which always matches any character other than
2131 newline. (It is the same as "." when PCRE_DOTALL is not set.)
2133 6. When compiling pcregrep with newer versions of gcc which may have
2134 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2135 declared with attribute warn_unused_result" were given. Just casting the
2136 result to (void) does not stop the warnings; a more elaborate fudge is
2137 needed. I've used a macro to implement this.
2139 7. Minor change to pcretest.c to avoid a compiler warning.
2141 8. Added four artifical Unicode properties to help with an option to make
2142 \s etc use properties (see next item). The new properties are: Xan
2143 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2145 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2146 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2147 this option. Modified pcretest to add /W to test this facility. Added
2148 REG_UCP to make it available via the POSIX interface.
2150 10. Added --line-buffered to pcregrep.
2152 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2153 studied, and the match started with a letter with a code point greater than
2154 127 whose first byte was different to the first byte of the other case of
2155 the letter, the other case of this starting letter was not recognized
2156 (#976).
2158 12. If a pattern that was studied started with a repeated Unicode property
2159 test, for example, \p{Nd}+, there was the theoretical possibility of
2160 setting up an incorrect bitmap of starting bytes, but fortunately it could
2161 not have actually happened in practice until change 8 above was made (it
2162 added property types that matched character-matching opcodes).
2164 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2165 possible starting bytes for non-anchored patterns.
2167 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2168 \R, and also a number of cases that involve Unicode properties, both
2169 explicit and implicit when PCRE_UCP is set.
2171 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2172 input, it could crash or give wrong results if characters with values
2173 greater than 0xc0 were present in the subject string. (Detail: it assumed
2174 UTF-8 input when processing these items.)
2176 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2177 size_t is 64-bit (#991).
2179 17. Added a check for running out of memory when PCRE is compiled with
2180 --disable-stack-for-recursion (#990).
2182 18. If the last data line in a file for pcretest does not have a newline on
2183 the end, a newline was missing in the output.
2185 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2186 less than 128) in its various bitmaps. However, there is a facility for
2187 generating tables according to the current locale when PCRE is compiled. It
2188 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2189 characters, are recognized by isspace() and therefore were getting set in
2190 these tables, and indeed these tables seem to approximate to ISO 8859. This
2191 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2192 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2193 which of course cannot start UTF-8 characters. I have changed the code so
2194 that only real ASCII characters (less than 128) and the correct starting
2195 bytes for UTF-8 encodings are set for characters greater than 127 when in
2196 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2197 altogether.)
2199 20. Added the /T option to pcretest so as to be able to run tests with non-
2200 standard character tables, thus making it possible to include the tests
2201 used for 19 above in the standard set of tests.
2203 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2204 reference to a subpattern the other side of a comment that contains an
2205 opening parenthesis caused either an internal compiling error, or a
2206 reference to the wrong subpattern.
2209 Version 8.02 19-Mar-2010
2210 ------------------------
2212 1. The Unicode data tables have been updated to Unicode 5.2.0.
2214 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2215 configured.
2217 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2218 original author of that file, following a query about its status.
2220 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2221 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2223 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2224 quantifier applied to a forward-referencing subroutine call, could compile
2225 incorrect code or give the error "internal error: previously-checked
2226 referenced subpattern not found".
2228 6. Both MS Visual Studio and Symbian OS have problems with initializing
2229 variables to point to external functions. For these systems, therefore,
2230 pcre_malloc etc. are now initialized to local functions that call the
2231 relevant global functions.
2233 7. There were two entries missing in the vectors called coptable and poptable
2234 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2235 I've fixed the data, and added a kludgy way of testing at compile time that
2236 the lengths are correct (equal to the number of opcodes).
2238 8. Following on from 7, I added a similar kludge to check the length of the
2239 eint vector in pcreposix.c.
2241 9. Error texts for pcre_compile() are held as one long string to avoid too
2242 much relocation at load time. To find a text, the string is searched,
2243 counting zeros. There was no check for running off the end of the string,
2244 which could happen if a new error number was added without updating the
2245 string.
2247 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2249 11. \K was not working if it appeared in an atomic group or in a group that
2250 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2251 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2252 the assertion is positive, but not if it is negative.
2254 12. Change 11 fortuitously reduced the size of the stack frame used in the
2255 "match()" function of pcre_exec.c by one pointer. Forthcoming
2256 implementation of support for (*MARK) will need an extra pointer on the
2257 stack; I have reserved it now, so that the stack frame size does not
2258 decrease.
2260 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2261 item in branch that calls a recursion is a subroutine call - as in the
2262 second branch in the above example - was incorrectly given the compile-
2263 time error "recursive call could loop indefinitely" because pcre_compile()
2264 was not correctly checking the subroutine for matching a non-empty string.
2266 14. The checks for overrunning compiling workspace could trigger after an
2267 overrun had occurred. This is a "should never occur" error, but it can be
2268 triggered by pathological patterns such as hundreds of nested parentheses.
2269 The checks now trigger 100 bytes before the end of the workspace.
2271 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2274 Version 8.01 19-Jan-2010
2275 ------------------------
2277 1. If a pattern contained a conditional subpattern with only one branch (in
2278 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2279 computed the wrong minimum data length (which is of course zero for such
2280 subpatterns). This could cause incorrect "no match" results.
2282 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2283 the pattern is reset in the first branch, pcre_compile() failed with
2284 "internal error: code overflow at offset...". This happened only when
2285 the reset was to the original external option setting. (An optimization
2286 abstracts leading options settings into an external setting, which was the
2287 cause of this.)
2289 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2290 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2291 assertion pattern did not match (meaning that the assertion was true), it
2292 was incorrectly treated as false if the SKIP had been reached during the
2293 matching. This also applied to assertions used as conditions.
2295 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2296 assertion subpattern, including such a pattern used as a condition,
2297 unpredictable results occurred, instead of the error return
2300 5. The C++ GlobalReplace function was not working like Perl for the special
2301 situation when an empty string is matched. It now does the fancy magic
2302 stuff that is necessary.
2304 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2305 removed. (These were left over from very, very early versions of PCRE.)
2307 7. Some cosmetic changes to the code to make life easier when compiling it
2308 as part of something else:
2310 (a) Change DEBUG to PCRE_DEBUG.
2312 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2313 called "current" as "current_branch", to prevent a collision with the
2314 Linux macro when compiled as a kernel module.
2316 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2317 prevent a collision with the Linux macro when compiled as a kernel
2318 module.
2320 8. In pcre_compile() there are some checks for integer overflows that used to
2321 cast potentially large values to (double). This has been changed to that
2322 when building, a check for int64_t is made, and if it is found, it is used
2323 instead, thus avoiding the use of floating point arithmetic. (There is no
2324 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2325 double.
2327 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2328 2005 (difference between two addresses compared to an unsigned value).
2330 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2331 custom one, because of the following reported problem in Windows:
2333 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2334 under Win32.
2335 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2336 therefore missing the function definition.
2337 - The compiler thus generates a "C" signature for the test function.
2338 - The linker fails to find the "C" function.
2339 - PCRE fails to configure if asked to do so against libbz2.
2341 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2342 messages were output:
2344 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2345 rerunning libtoolize, to keep the correct libtool macros in-tree.
2346 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2348 I have done both of these things.
2350 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2351 most of the time, it *can* run out if it is given a pattern that contains a
2352 runaway infinite recursion. I updated the discussion in the pcrestack man
2353 page.
2355 13. Now that we have gone to the x.xx style of version numbers, the minor
2356 version may start with zero. Using 08 or 09 is a bad idea because users
2357 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2358 interpreted as invalid octal numbers. I've updated the previous comment in
2359 configure.ac, and also added a check that gives an error if 08 or 09 are
2360 used.
2362 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2363 causing partial matching to fail when the end of the subject matched \W
2364 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2366 15. There were some discrepancies between the declarations in pcre_internal.h
2367 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2368 their definitions. The declarations used "const uschar *" and the
2369 definitions used USPTR. Even though USPTR is normally defined as "const
2370 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2371 reported that: "This difference in casting confuses some C++ compilers, for
2372 example, SunCC recognizes above declarations as different functions and
2373 generates broken code for hbpcre." I have changed the declarations to use
2374 USPTR.
2376 16. GNU libtool is named differently on some systems. The autogen.sh script now
2377 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2378 (FreeBSD).
2380 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2381 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2382 comment: "Figure out how to create a longlong from a string: strtoll and
2383 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2384 instance, but it only takes 2 args instead of 3!"
2386 18. A subtle bug concerned with back references has been fixed by a change of
2387 specification, with a corresponding code fix. A pattern such as
2388 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2389 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2390 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2391 same bug. Such groups have to be quantified to be useful, or contained
2392 inside another quantified group. (If there's no repetition, the reference
2393 can never match.) The problem arises because, having left the group and
2394 moved on to the rest of the pattern, a later failure that backtracks into
2395 the group uses the captured value from the final iteration of the group
2396 rather than the correct earlier one. I have fixed this in PCRE by forcing
2397 any group that contains a reference to itself to be an atomic group; that
2398 is, there cannot be any backtracking into it once it has completed. This is
2399 similar to recursive and subroutine calls.
2402 Version 8.00 19-Oct-09
2403 ----------------------
2405 1. The table for translating pcre_compile() error codes into POSIX error codes
2406 was out-of-date, and there was no check on the pcre_compile() error code
2407 being within the table. This could lead to an OK return being given in
2408 error.
2410 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2411 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2412 in a Windows environment.
2414 3. The pcregrep --count option prints the count for each file even when it is
2415 zero, as does GNU grep. However, pcregrep was also printing all files when
2416 --files-with-matches was added. Now, when both options are given, it prints
2417 counts only for those files that have at least one match. (GNU grep just
2418 prints the file name in this circumstance, but including the count seems
2419 more useful - otherwise, why use --count?) Also ensured that the
2420 combination -clh just lists non-zero counts, with no names.
2422 4. The long form of the pcregrep -F option was incorrectly implemented as
2423 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2424 but it seems right to fix it, and I didn't think it was worth preserving
2425 the old behaviour.
2427 5. The command line items --regex=pattern and --regexp=pattern were not
2428 recognized by pcregrep, which required --regex pattern or --regexp pattern
2429 (with a space rather than an '='). The man page documented the '=' forms,
2430 which are compatible with GNU grep; these now work.
2432 6. No libpcreposix.pc file was created for pkg-config; there was just
2433 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2435 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2436 when UCP support is not needed, by modifying the Python script that
2437 generates it from Unicode data files. This should not matter if the module
2438 is correctly used as a library, but I received one complaint about 50K of
2439 unwanted data. My guess is that the person linked everything into his
2440 program rather than using a library. Anyway, it does no harm.
2442 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2443 was a minimum greater than 1 for a wide character in a possessive
2444 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2445 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2446 character. Chaos in the form of incorrect output or a compiling loop could
2447 result.
2449 9. The restrictions on what a pattern can contain when partial matching is
2450 requested for pcre_exec() have been removed. All patterns can now be
2451 partially matched by this function. In addition, if there are at least two
2452 slots in the offset vector, the offset of the earliest inspected character
2453 for the match and the offset of the end of the subject are set in them when
2454 PCRE_ERROR_PARTIAL is returned.
2456 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2457 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2458 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2459 and may be more useful for multi-segment matching.
2461 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2462 used to be given if ever the end of the subject was reached; now it is
2463 given only if matching could not proceed because another character was
2464 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2465 string "Z", which now yields "no match" instead of "partial match". In the
2466 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2467 final character ended with (*FAIL).
2469 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2470 if the pattern had a "must contain" character that was already found in the
2471 earlier partial match, unless partial matching was again requested. For
2472 example, with the pattern /dog.(body)?/, the "must contain" character is
2473 "g". If the first part-match was for the string "dog", restarting with
2474 "sbody" failed. This bug has been fixed.
2476 13. The string returned by pcre_dfa_exec() after a partial match has been
2477 changed so that it starts at the first inspected character rather than the
2478 first character of the match. This makes a difference only if the pattern
2479 starts with a lookbehind assertion or \b or \B (\K is not supported by
2480 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2481 matching functions compatible, and I think it's the right thing to do.
2483 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2484 so that the demonstration program is easily available in environments where
2485 PCRE has not been installed from source.
2487 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2488 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2489 library.
2491 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2492 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2493 is not the first non-POSIX option to be added. Clearly some people find
2494 these options useful.
2496 17. If a caller to the POSIX matching function regexec() passes a non-zero
2497 value for nmatch with a NULL value for pmatch, the value of
2498 nmatch is forced to zero.
2500 18. RunGrepTest did not have a test for the availability of the -u option of
2501 the diff command, as RunTest does. It now checks in the same way as
2502 RunTest, and also checks for the -b option.
2504 19. If an odd number of negated classes containing just a single character
2505 interposed, within parentheses, between a forward reference to a named
2506 subpattern and the definition of the subpattern, compilation crashed with
2507 an internal error, complaining that it could not find the referenced
2508 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2509 [The bug was that it was starting one character too far in when skipping
2510 over the character class, thus treating the ] as data rather than
2511 terminating the class. This meant it could skip too much.]
2513 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2514 /g option in pcretest when the pattern contains \K, which makes it possible
2515 to have an empty string match not at the start, even when the pattern is
2516 anchored. Updated pcretest and pcredemo to use this option.
2518 21. If the maximum number of capturing subpatterns in a recursion was greater
2519 than the maximum at the outer level, the higher number was returned, but
2520 with unset values at the outer level. The correct (outer level) value is
2521 now given.
2523 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2524 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2525 make it do so. The string so far is captured, making this feature
2526 compatible with Perl.
2528 23. The tests have been re-organized, adding tests 11 and 12, to make it
2529 possible to check the Perl 5.10 features against Perl 5.10.
2531 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2532 pattern matches a fixed length string. PCRE did not allow this; now it
2533 does. Neither allows recursion.
2535 25. I finally figured out how to implement a request to provide the minimum
2536 length of subject string that was needed in order to match a given pattern.
2537 (It was back references and recursion that I had previously got hung up
2538 on.) This code has now been added to pcre_study(); it finds a lower bound
2539 to the length of subject needed. It is not necessarily the greatest lower
2540 bound, but using it to avoid searching strings that are too short does give
2541 some useful speed-ups. The value is available to calling programs via
2542 pcre_fullinfo().
2544 26. While implementing 25, I discovered to my embarrassment that pcretest had
2545 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2546 study optimizations had never been tested with that matching function.
2547 Oops. What is worse, even when it was passed study data, there was a bug in
2548 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2549 were also very few tests of studied patterns with pcre_dfa_exec().
2551 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2552 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2553 on the other side of the coin, they are no longer allowed to have different
2554 names, because these cannot be distinguished in PCRE, and this has caused
2555 confusion. (This is a difference from Perl.)
2557 28. When duplicate subpattern names are present (necessarily with different
2558 numbers, as required by 27 above), and a test is made by name in a
2559 conditional pattern, either for a subpattern having been matched, or for
2560 recursion in such a pattern, all the associated numbered subpatterns are
2561 tested, and the overall condition is true if the condition is true for any
2562 one of them. This is the way Perl works, and is also more like the way
2563 testing by number works.
2566 Version 7.9 11-Apr-09
2567 ---------------------
2569 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2570 (pcretest), all targets were linked against these libraries. This included
2571 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2572 libraries. This caused unwanted dependencies to be created. This problem
2573 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2574 pcretest is linked with readline.
2576 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2577 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2578 moved outside it again, because FALSE and TRUE are already defined in AIX,
2579 but BOOL is not.
2581 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2582 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2584 4. The pcregrep documentation said spaces were inserted as well as colons (or
2585 hyphens) following file names and line numbers when outputting matching
2586 lines. This is not true; no spaces are inserted. I have also clarified the
2587 wording for the --colour (or --color) option.
2589 5. In pcregrep, when --colour was used with -o, the list of matching strings
2590 was not coloured; this is different to GNU grep, so I have changed it to be
2591 the same.
2593 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2594 each matching line was coloured. Now it goes on to look for further matches
2595 of any of the test patterns, which is the same behaviour as GNU grep.
2597 7. A pattern that could match an empty string could cause pcregrep to loop; it
2598 doesn't make sense to accept an empty string match in pcregrep, so I have
2599 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2600 seems to be how GNU grep behaves. [But see later change 40 for release
2601 8.33.]
2603 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2604 start or after a newline", because the conditional assertion was not being
2605 correctly handled. The rule now is that both the assertion and what follows
2606 in the first alternative must satisfy the test.
2608 9. If auto-callout was enabled in a pattern with a conditional group whose
2609 condition was an assertion, PCRE could crash during matching, both with
2610 pcre_exec() and pcre_dfa_exec().
2612 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2613 used for matching.
2615 11. Unicode property support in character classes was not working for
2616 characters (bytes) greater than 127 when not in UTF-8 mode.
2618 12. Added the -M command line option to pcretest.
2620 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2622 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2624 16. Added comments and documentation about mis-use of no_arg in the C++
2625 wrapper.
2627 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2628 from Martin Jerabek that uses macro names for all relevant character and
2629 string constants.
2631 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2632 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2633 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2634 these, but not everybody uses configure.
2636 19. A conditional group that had only one branch was not being correctly
2637 recognized as an item that could match an empty string. This meant that an
2638 enclosing group might also not be so recognized, causing infinite looping
2639 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2640 with the subject "ab", where knowledge that the repeated group can match
2641 nothing is needed in order to break the loop.
2643 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2644 exec(), but without supplying a callout function, matching went wrong.
2646 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2647 leak if the size of the offset vector was greater than 30. When the vector
2648 is smaller, the saved offsets during recursion go onto a local stack
2649 vector, but for larger vectors malloc() is used. It was failing to free
2650 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2651 error, in fact).
2653 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2654 heapframe that is used only when UTF-8 support is enabled. This caused no
2655 problem, but was untidy.
2657 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2658 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2659 included within another project.
2661 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2662 slightly modified by me:
2664 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2665 not building pcregrep.
2667 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2668 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2670 25. Forward references, both numeric and by name, in patterns that made use of
2671 duplicate group numbers, could behave incorrectly or give incorrect errors,
2672 because when scanning forward to find the reference group, PCRE was not
2673 taking into account the duplicate group numbers. A pattern such as
2674 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2676 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2677 the feature of a custom pointer more persuasive (as requested by a user).
2679 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2680 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2681 pre-defined.
2683 28. Added support for (*UTF8) at the start of a pattern.
2685 29. Arrange for flags added by the "release type" setting in CMake to be shown
2686 in the configuration summary.
2689 Version 7.8 05-Sep-08
2690 ---------------------
2692 1. Replaced UCP searching code with optimized version as implemented for Ad
2693 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2694 stage table and inline lookup instead of a function, giving speed ups of 2
2695 to 5 times on some simple patterns that I tested. Permission was given to
2696 distribute the MultiStage2.py script that generates the tables (it's not in
2697 the tarball, but is in the Subversion repository).
2699 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2700 scripts.
2702 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2703 a group with a zero qualifier. The result of the study could be incorrect,
2704 or the function might crash, depending on the pattern.
2706 4. Caseless matching was not working for non-ASCII characters in back
2707 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2708 It now works when Unicode Property Support is available.
2710 5. In pcretest, an escape such as \x{de} in the data was always generating
2711 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2712 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2713 truncation.
2715 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2717 7. Added two (int) casts to pcregrep when printing the difference of two
2718 pointers, in case they are 64-bit values.
2720 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2721 test 2 if it fails.
2723 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2724 and a #define of that name to empty if it is not externally set. This is to
2725 allow users of MSVC to set it if necessary.
2727 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2728 the convenience functions in the pcre_get.c source file.
2730 11. An option change at the start of a pattern that had top-level alternatives
2731 could cause overwriting and/or a crash. This command provoked a crash in
2732 some environments:
2734 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2736 This potential security problem was recorded as CVE-2008-2371.
2738 12. For a pattern where the match had to start at the beginning or immediately
2739 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2740 pcre_dfa_exec() could read past the end of the passed subject if there was
2741 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2742 pcretest so that it places the subject at the end of its malloc-ed buffer.
2744 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2745 exec() might read past the end of the data buffer in UTF-8 mode.
2747 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2748 the data contained the byte 0x85 as part of a UTF-8 character within its
2749 first line. This applied both to normal and DFA matching.
2751 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2752 /^[^d]*?$/8 failed to match "abc".
2754 16. Added a missing copyright notice to pcrecpp_internal.h.
2756 17. Make it more clear in the documentation that values returned from
2757 pcre_exec() in ovector are byte offsets, not character counts.
2759 18. Tidied a few places to stop certain compilers from issuing warnings.
2761 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2762 supplied by Stefan Weber. I made a further small update for 7.8 because
2763 there is a change of source arrangements: the pcre_searchfuncs.c module is
2764 replaced by pcre_ucd.c.
2767 Version 7.7 07-May-08
2768 ---------------------
2770 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2771 a string to a long long, pretend we don't even have a long long." This is
2772 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2774 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2775 pre-7.6 versions, which defined a global no_arg variable instead of putting
2776 it in the RE class. (See also #8 below.)
2778 3. Remove a line of dead code, identified by coverity and reported by Nuno
2779 Lopes.
2781 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2783 (1) The include/exclude patterns were being applied to the whole pathnames
2784 of files, instead of just to the final components.
2786 (2) If there was more than one level of directory, the subdirectories were
2787 skipped unless they satisfied the include/exclude conditions. This is
2788 inconsistent with GNU grep (and could even be seen as contrary to the
2789 pcregrep specification - which I improved to make it absolutely clear).
2790 The action now is always to scan all levels of directory, and just
2791 apply the include/exclude patterns to regular files.
2793 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2794 --exclude_dir in the tests to avoid scanning .svn directories.
2796 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2797 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2798 doesn't support NULs in patterns.
2800 7. Added some missing "const"s to declarations of static tables in
2801 pcre_compile.c and pcre_dfa_exec.c.
2803 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2804 caused by fix #2 above. (Subsequently also a second patch to fix the
2805 first patch. And a third patch - this was a messy problem.)
2807 9. Applied Craig's patch to remove the use of push_back().
2809 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2810 matching function regexec().
2812 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2813 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2814 references. PCRE supports relative numbers with this syntax (I don't think
2815 Oniguruma does).
2817 12. Previously, a group with a zero repeat such as (...){0} was completely
2818 omitted from the compiled regex. However, this means that if the group
2819 was called as a subroutine from elsewhere in the pattern, things went wrong
2820 (an internal error was given). Such groups are now left in the compiled
2821 pattern, with a new opcode that causes them to be skipped at execution
2822 time.
2824 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2825 to the way PCRE behaves:
2827 (a) A lone ] character is dis-allowed (Perl treats it as data).
2829 (b) A back reference to an unmatched subpattern matches an empty string
2830 (Perl fails the current match path).
2832 (c) A data ] in a character class must be notated as \] because if the
2833 first data character in a class is ], it defines an empty class. (In
2834 Perl it is not possible to have an empty class.) The empty class []
2835 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2836 The negative empty class [^] matches any one character, independently
2837 of the DOTALL setting.
2839 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2840 non-existent subpattern following a character class starting with ']' and
2841 containing () gave an internal compiling error instead of "reference to
2842 non-existent subpattern". Fortunately, when the pattern did exist, the
2843 compiled code was correct. (When scanning forwards to check for the
2844 existence of the subpattern, it was treating the data ']' as terminating
2845 the class, so got the count wrong. When actually compiling, the reference
2846 was subsequently set up correctly.)
2848 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2849 it was being rejected as not supported by pcre_dfa_exec(), even though
2850 other assertions are supported. I have made pcre_dfa_exec() support
2851 (*FAIL).
2853 16. The implementation of 13c above involved the invention of a new opcode,
2854 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2855 cannot be changed at match time, I realized I could make a small
2856 improvement to matching performance by compiling OP_ALLANY instead of
2857 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2858 on the OP_ANY path.
2860 17. Compiling pcretest on Windows with readline support failed without the
2861 following two fixes: (1) Make the unistd.h include conditional on
2862 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2864 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2865 ncurses library to be included for pcretest when ReadLine support is
2866 requested, but also to allow for it to be overridden. This patch came from
2867 Daniel Bergström.
2869 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2870 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2871 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2872 spotting this.
2875 Version 7.6 28-Jan-08
2876 ---------------------
2878 1. A character class containing a very large number of characters with
2879 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2880 overflow.
2882 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2883 HAVE_LONG_LONG is not defined.
2885 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2886 bring them up to date and include new features. This patch includes:
2888 - Fixed PH's badly added libz and libbz2 support.
2889 - Fixed a problem with static linking.
2890 - Added pcredemo. [But later removed - see 7 below.]
2891 - Fixed dftables problem and added an option.
2892 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2894 - Added readline support for pcretest.
2895 - Added an listing of the option settings after cmake has run.
2897 4. A user submitted a patch to Makefile that makes it easy to create
2898 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2899 Makefile.am that cause it to include this special target, without
2900 affecting anything else. Note that the same mingw target plus all
2901 the other distribution libraries and programs are now supported
2902 when configuring with CMake (see 6 below) instead of with
2903 Configure/Make.
2905 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2906 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2907 exported in the Windows port". It has not yet been confirmed that the patch
2908 solves the problem, but it does no harm.
2910 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2911 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2912 with CMake, and also correct the comment about stack recursion.
2914 7. Remove the automatic building of pcredemo from the ./configure system and
2915 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2916 of a program that users should build themselves after PCRE is installed, so
2917 building it automatically is not really right. What is more, it gave
2918 trouble in some build environments.
2920 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2923 Version 7.5 10-Jan-08
2924 ---------------------
2926 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2927 values in parens when parsing an RE using the C++ wrapper."
2929 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2930 Characters greater than 255 were excluded from the class instead of being
2931 included.
2933 3. The same bug as (2) above applied to negated POSIX classes such as
2934 [:^space:].
2936 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2937 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2938 I have changed it.
2940 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2941 first named subpattern) and a construct such as (?&a) would reference the
2942 first named subpattern whose name started with "a" (in other words, the
2943 length check was missing). Both these problems are fixed. "Subpattern name
2944 expected" is now given for (?&) (a zero-length name), and this patch also
2945 makes it give the same error for \k'' (previously it complained that that
2946 was a reference to a non-existent subpattern).
2948 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2949 this is right because (?- can be followed by option settings as well as by
2950 digits. I have, however, made the messages clearer.
2952 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2953 than the number used in the conditional) now cause a compile-time error.
2954 This is actually not compatible with Perl, which accepts such patterns, but
2955 treats the conditional as always being FALSE (as PCRE used to), but it
2956 seems to me that giving a diagnostic is better.
2958 8. Change "alphameric" to the more common word "alphanumeric" in comments
2959 and messages.
2961 9. Fix two occurrences of "backslash" in comments that should have been
2962 "backspace".
2964 10. Remove two redundant lines of code that can never be obeyed (their function
2965 was moved elsewhere).
2967 11. The program that makes PCRE's Unicode character property table had a bug
2968 which caused it to generate incorrect table entries for sequences of
2969 characters that have the same character type, but are in different scripts.
2970 It amalgamated them into a single range, with the script of the first of
2971 them. In other words, some characters were in the wrong script. There were
2972 thirteen such cases, affecting characters in the following ranges:
2974 U+002b0 - U+002c1
2975 U+0060c - U+0060d
2976 U+0061e - U+00612
2977 U+0064b - U+0065e
2978 U+0074d - U+0076d
2979 U+01800 - U+01805
2980 U+01d00 - U+01d77
2981 U+01d9b - U+01dbf
2982 U+0200b - U+0200f
2983 U+030fc - U+030fe
2984 U+03260 - U+0327f
2985 U+0fb46 - U+0fbb1
2986 U+10450 - U+1049d
2988 12. The -o option (show only the matching part of a line) for pcregrep was not
2989 compatible with GNU grep in that, if there was more than one match in a
2990 line, it showed only the first of them. It now behaves in the same way as
2991 GNU grep.
2993 13. If the -o and -v options were combined for pcregrep, it printed a blank
2994 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2995 does the same. The return code can be used to tell if there were any
2996 non-matching lines.
2998 14. Added --file-offsets and --line-offsets to pcregrep.
3000 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
3001 infinitely looping recursion. The bug was that positive lookaheads were not
3002 being skipped when checking for a possible empty match (negative lookaheads
3003 and both kinds of lookbehind were skipped).
3005 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
3006 inclusion of <windows.h> to before rather than after the definition of
3007 INVALID_FILE_ATTRIBUTES (patch from David Byron).
3009 17. Specifying a possessive quantifier with a specific limit for a Unicode
3010 character property caused pcre_compile() to compile bad code, which led at
3011 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
3012 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
3013 caused the error; without that there was no problem.
3015 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
3017 19. Added --enable-pcretest-libreadline.
3019 20. In pcrecpp.cc, the variable 'count' was incremented twice in
3020 RE::GlobalReplace(). As a result, the number of replacements returned was
3021 double what it should be. I removed one of the increments, but Craig sent a
3022 later patch that removed the other one (the right fix) and added unit tests
3023 that check the return values (which was not done before).
3025 21. Several CMake things:
3027 (1) Arranged that, when cmake is used on Unix, the libraries end up with
3028 the names libpcre and libpcreposix, not just pcre and pcreposix.
3030 (2) The above change means that pcretest and pcregrep are now correctly
3031 linked with the newly-built libraries, not previously installed ones.
3035 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
3036 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
3037 UTF-8 newline character). The key issue is that the pattern starts .*;
3038 this means that the match must be either at the beginning, or after a
3039 newline. The bug was in the code for advancing after a failed match and
3040 checking that the new position followed a newline. It was not taking
3041 account of UTF-8 characters correctly.
3043 23. PCRE was behaving differently from Perl in the way it recognized POSIX
3044 character classes. PCRE was not treating the sequence [:...:] as a
3045 character class unless the ... were all letters. Perl, however, seems to
3046 allow any characters between [: and :], though of course it rejects as
3047 unknown any "names" that contain non-letters, because all the known class
3048 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
3049 for example, whereas PCRE did not - it did not recognize a POSIX character
3050 class. This seemed a bit dangerous, so the code has been changed to be
3051 closer to Perl. The behaviour is not identical to Perl, because PCRE will
3052 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
3053 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
3054 Perl does, and where it didn't before.
3056 24. Rewrite so as to remove the single use of %n from pcregrep because in some
3057 Windows environments %n is disabled by default.
3060 Version 7.4 21-Sep-07
3061 ---------------------
3063 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
3064 means that a class such as [\s] counted as "explicit reference to CR or
3065 LF". That isn't really right - the whole point of the change was to try to
3066 help when there was an actual mention of one of the two characters. So now
3067 the change happens only if \r or \n (or a literal CR or LF) character is
3068 encountered.
3070 2. The 32-bit options word was also used for 6 internal flags, but the numbers
3071 of both had grown to the point where there were only 3 bits left.
3072 Fortunately, there was spare space in the data structure, and so I have
3073 moved the internal flags into a new 16-bit field to free up more option
3074 bits.
3076 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3077 but did not set the internal JCHANGED flag - either of these is enough to
3078 control the way the "get" function works - but the PCRE_INFO_JCHANGED
3079 facility is supposed to tell if (?J) was ever used, so now (?J) at the
3080 start sets both bits.
3082 4. Added options (at build time, compile time, exec time) to change \R from
3083 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3085 5. doc/pcresyntax.html was missing from the distribution.
3087 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3088 compatibility, even though it is no longer used.
3090 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3091 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3092 windows.h file is present (where different names are used). [This was
3093 reversed later after testing - see 16 below.]
3095 8. Changed all #include <config.h> to #include "config.h". There were also
3096 some further <pcre.h> cases that I changed to "pcre.h".
3098 9. When pcregrep was used with the --colour option, it missed the line ending
3099 sequence off the lines that it output.
3101 10. It was pointed out to me that arrays of string pointers cause lots of
3102 relocations when a shared library is dynamically loaded. A technique of
3103 using a single long string with a table of offsets can drastically reduce
3104 these. I have refactored PCRE in four places to do this. The result is
3105 dramatic:
3107 Originally: 290
3108 After changing UCP table: 187
3109 After changing error message table: 43
3110 After changing table of "verbs" 36
3111 After changing table of Posix names 22
3113 Thanks to the folks working on Gregex for glib for this insight.
3115 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3116 unicode-properties was also set.
3118 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3120 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3121 checked only for CRLF.
3123 14. Added casts to pcretest.c to avoid compiler warnings.
3125 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3127 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3128 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3129 entirely. This removes changes made in 7 above.
3131 17. The CMake files have been updated, and there is now more information about
3132 building with CMake in the NON-UNIX-USE document.
3135 Version 7.3 28-Aug-07
3136 ---------------------
3138 1. In the rejigging of the build system that eventually resulted in 7.1, the
3139 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3140 brackets there is not right, since it causes compilers to look for an
3141 installed pcre.h, not the version that is in the source that is being
3142 compiled (which of course may be different). I have changed it back to:
3144 #include "pcre.h"
3146 I have a vague recollection that the change was concerned with compiling in
3147 different directories, but in the new build system, that is taken care of
3148 by the VPATH setting the Makefile.
3150 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3151 when the subject happened to end in the byte 0x85 (e.g. if the last
3152 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3153 characters but of course it shouldn't be taken as a newline when it is part
3154 of another character. The bug was that, for an unlimited repeat of . in
3155 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3156 characters when looking for a newline.
3158 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3160 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3161 in debug output.
3163 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3164 long printing in the pcrecpp unittest when running under MinGW.
3166 6. ESC_K was left out of the EBCDIC table.
3168 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3169 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3170 limit also applies to "virtual nesting" when a pattern is recursive, and in
3171 this case 1000 isn't so big. I have been able to remove this limit at the
3172 expense of backing off one optimization in certain circumstances. Normally,
3173 when pcre_exec() would call its internal match() function recursively and
3174 immediately return the result unconditionally, it uses a "tail recursion"
3175 feature to save stack. However, when a subpattern that can match an empty
3176 string has an unlimited repetition quantifier, it no longer makes this
3177 optimization. That gives it a stack frame in which to save the data for
3178 checking that an empty string has been matched. Previously this was taken
3179 from the 1000-entry workspace that had been reserved. So now there is no
3180 explicit limit, but more stack is used.
3182 8. Applied Daniel's patches to solve problems with the import/export magic
3183 syntax that is required for Windows, and which was going wrong for the
3184 pcreposix and pcrecpp parts of the library. These were overlooked when this
3185 problem was solved for the main library.
3187 9. There were some crude static tests to avoid integer overflow when computing
3188 the size of patterns that contain repeated groups with explicit upper
3189 limits. As the maximum quantifier is 65535, the maximum group length was
3190 set at 30,000 so that the product of these two numbers did not overflow a
3191 32-bit integer. However, it turns out that people want to use groups that
3192 are longer than 30,000 bytes (though not repeat them that many times).
3193 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3194 made it possible to implement the integer overflow checks in a much more
3195 dynamic way, which I have now done. The artificial limitation on group
3196 length has been removed - we now have only the limit on the total length of
3197 the compiled pattern, which depends on the LINK_SIZE setting.
3199 10. Fixed a bug in the documentation for get/copy named substring when
3200 duplicate names are permitted. If none of the named substrings are set, the
3201 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3202 empty string.
3204 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3205 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3206 because the ] is interpreted as the first data character and the
3207 terminating ] is not found. PCRE has been made compatible with Perl in this
3208 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3209 cause memory overwriting.
3211 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3212 string has been matched (to stop an infinite loop). It was not recognizing
3213 a conditional subpattern that could match an empty string if that
3214 subpattern was within another subpattern. For example, it looped when
3215 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3216 condition was not nested. This bug has been fixed.
3218 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3219 past the start of the subject in the presence of bytes with the top bit
3220 set, for example "\x8aBCD".
3222 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3223 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3225 14. Optimized (?!) to (*FAIL).
3227 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3228 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3229 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3230 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3231 does: it's just the validity check that is more restrictive.
3233 16. Inserted checks for integer overflows during escape sequence (backslash)
3234 processing, and also fixed erroneous offset values for syntax errors during
3235 backslash processing.
3237 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3238 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3240 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3241 caused an overrun.
3243 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3244 something other than just ASCII characters) inside a group that had an
3245 unlimited repeat caused a loop at compile time (while checking to see
3246 whether the group could match an empty string).
3248 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3249 [\P{Any}] did so. (Error in the code for printing property names.)
3251 21. An orphan \E inside a character class could cause a crash.
3253 22. A repeated capturing bracket such as (A)? could cause a wild memory
3254 reference during compilation.
3256 23. There are several functions in pcre_compile() that scan along a compiled
3257 expression for various reasons (e.g. to see if it's fixed length for look
3258 behind). There were bugs in these functions when a repeated \p or \P was
3259 present in the pattern. These operators have additional parameters compared
3260 with \d, etc, and these were not being taken into account when moving along
3261 the compiled data. Specifically:
3263 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3264 length.
3266 (b) An item such as \pL+ within a repeated group could cause crashes or
3267 loops.
3269 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3270 "reference to non-existent subpattern" error.
3272 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3274 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3275 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3277 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3279 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3280 character were causing crashes (broken optimization).
3282 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3283 \p or \P) caused a compile-time loop.
3285 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3286 break. For example, the unstudied pattern [\r\n]A does not match the string
3287 "\r\nA" because change 7.0/46 below moves the current point on by two
3288 characters after failing to match at the start. However, the pattern \nA
3289 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3290 the same is true. There doesn't seem any very clean way out of this, but
3291 what I have chosen to do makes the common cases work: PCRE now takes note
3292 of whether there can be an explicit match for \r or \n anywhere in the
3293 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3294 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3295 pattern has explicit CR or LF references.
3297 29. Added (*CR) etc for changing newline setting at start of pattern.
3300 Version 7.2 19-Jun-07
3301 ---------------------
3303 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3304 which is apparently normally available under Windows.
3306 2. Re-jig the pcregrep tests with different newline settings in an attempt
3307 to make them independent of the local environment's newline setting.
3309 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3311 4. Some of the "internals" tests were previously cut out when the link size
3312 was not 2, because the output contained actual offsets. The recent new
3313 "Z" feature of pcretest means that these can be cut out, making the tests
3314 usable with all link sizes.
3316 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3317 stack recursion. This gives a massive performance boost under BSD, but just
3318 a small improvement under Linux. However, it saves one field in the frame
3319 in all cases.
3321 6. Added more features from the forthcoming Perl 5.10:
3323 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3324 recursion call. It refers to the nth most recently opened parentheses.
3326 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3327 to be opened parentheses.
3329 (c) Conditions that refer to capturing parentheses can be specified
3330 relatively, for example, (?(-2)... or (?(+3)...
3332 (d) \K resets the start of the current match so that everything before
3333 is not part of it.
3335 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3337 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3338 reference syntax.
3340 (g) (?| introduces a group in which the numbering of parentheses in each
3341 alternative starts with the same number.
3343 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3345 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3348 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3349 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3350 for detecting groups that can match an empty string.
3352 9. A pattern with a very large number of alternatives (more than several
3353 hundred) was running out of internal workspace during the pre-compile
3354 phase, where pcre_compile() figures out how much memory will be needed. A
3355 bit of new cunning has reduced the workspace needed for groups with
3356 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3357 workspace instead of running out of the 4096 that are available.
3359 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3361 11. Applied patch from Google to remove an optimization that didn't quite work.
3362 The report of the bug said:
3364 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3365 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3366 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3368 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3369 it matched the wrong number of bytes.
3372 Version 7.1 24-Apr-07
3373 ---------------------
3375 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3376 that is more "standard", making use of automake and other Autotools. There
3377 is some re-arrangement of the files and adjustment of comments consequent
3378 on this.
3380 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3381 for recursive directory scanning broke on some systems because the files
3382 are not scanned in any specific order and on different systems the order
3383 was different. A call to "sort" has been inserted into RunGrepTest for the
3384 approprate test as a short-term fix. In the longer term there may be an
3385 alternative.
3387 3. I had an email from Eric Raymond about problems translating some of PCRE's
3388 man pages to HTML (despite the fact that I distribute HTML pages, some
3389 people do their own conversions for various reasons). The problems
3390 concerned the use of low-level troff macros .br and .in. I have therefore
3391 removed all such uses from the man pages (some were redundant, some could
3392 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3393 HTML has been updated to handle .nf/.fi and to complain if it encounters
3394 .br or .in.
3396 4. Updated comments in configure.ac that get placed in config.h.in and also
3397 arranged for config.h to be included in the distribution, with the name
3398 config.h.generic, for the benefit of those who have to compile without
3399 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3401 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3402 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3403 makevp.bat for the current PCRE, using the additional files
3404 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3406 6. A Windows user reported a minor discrepancy with test 2, which turned out
3407 to be caused by a trailing space on an input line that had got lost in his
3408 copy. The trailing space was an accident, so I've just removed it.
3410 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3411 that is needed.
3413 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3414 as "const" (a) because they are and (b) because it helps the PHP
3415 maintainers who have recently made a script to detect big data structures
3416 in the php code that should be moved to the .rodata section. I remembered
3417 to update Builducptable as well, so it won't revert if ucptable.h is ever
3418 re-created.
3420 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3421 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3422 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3423 support is not required. This saves 1.5-2K of code, which is important in
3424 some applications.
3426 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3427 so as not to refer to the tables, even though these functions will never be
3428 called when UTF-8 support is disabled. Otherwise there are problems with a
3429 shared library.
3431 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3433 (a) It was defining its arguments as char * instead of void *.
3435 (b) It was assuming that all moves were upwards in memory; this was true
3436 a long time ago when I wrote it, but is no longer the case.
3438 The emulated memove() is provided for those environments that have neither
3439 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3440 is clearly not the case, as these two bugs were recently reported.
3442 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3443 and Detrail to create the HTML documentation, the .txt form of the man
3444 pages, and it removes trailing spaces from listed files. It also creates
3445 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3446 case, it wraps all the #defines with #ifndefs. This script should be run
3447 before "make dist".
3449 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3450 with Unicode property support.
3452 (a) For a maximizing quantifier, if the two different cases of the
3453 character were of different lengths in their UTF-8 codings (there are
3454 some cases like this - I found 11), and the matching function had to
3455 back up over a mixture of the two cases, it incorrectly assumed they
3456 were both the same length.
3458 (b) When PCRE was configured to use the heap rather than the stack for
3459 recursion during matching, it was not correctly preserving the data for
3460 the other case of a UTF-8 character when checking ahead for a match
3461 while processing a minimizing repeat. If the check also involved
3462 matching a wide character, but failed, corruption could cause an
3463 erroneous result when trying to check for a repeat of the original
3464 character.
3466 13. Some tidying changes to the testing mechanism:
3468 (a) The RunTest script now detects the internal link size and whether there
3469 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3470 values substituted by "configure". (The RunGrepTest script already did
3471 this for UTF-8.) The configure.ac script no longer substitutes the
3472 relevant variables.
3474 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3475 with length and offset values. This means that the output is different
3476 for different internal link sizes. Test 2 is skipped for link sizes
3477 other than 2 because of this, bypassing the problem. Unfortunately,
3478 there was also a test in test 3 (the locale tests) that used /B and
3479 failed for link sizes other than 2. Rather than cut the whole test out,
3480 I have added a new /Z option to pcretest that replaces the length and
3481 offset values with spaces. This is now used to make test 3 independent
3482 of link size. (Test 2 will be tidied up later.)
3484 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3485 segmentation fault instead of returning the appropriate error message.
3487 15. In multiline mode when the newline sequence was set to "any", the pattern
3488 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3489 This doesn't seem right; it now treats the CRLF combination as the line
3490 ending, and so does not match in that case. It's only a pattern such as ^$
3491 that would hit this one: something like ^ABC$ would have failed after \r
3492 and then tried again after \r\n.
3494 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3495 in an attempt to make files that differ only in their line terminators
3496 compare equal. This works on Linux.
3498 17. Under certain error circumstances pcregrep might try to free random memory
3499 as it exited. This is now fixed, thanks to valgrind.
3501 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3502 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3503 was because its rules for how to advance for /g after matching an empty
3504 string at the end of a line did not allow for this case. They now check for
3505 it specially.
3507 20. pcretest is supposed to handle patterns and data of any length, by
3508 extending its buffers when necessary. It was getting this wrong when the
3509 buffer for a data line had to be extended.
3511 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3512 CRLF as a newline sequence.
3514 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3515 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3516 I have nevertheless tidied it up.
3518 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3520 24. Added a man page for pcre-config.
3523 Version 7.0 19-Dec-06
3524 ---------------------
3526 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3527 moving to gcc 4.1.1.
3529 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3530 sys/time.h, which is documented as needed for this function. It doesn't
3531 seem to matter on Linux, but it showed up on some releases of OS X.
3533 3. It seems that there are systems where bytes whose values are greater than
3534 127 match isprint() in the "C" locale. The "C" locale should be the
3535 default when a C program starts up. In most systems, only ASCII printing
3536 characters match isprint(). This difference caused the output from pcretest
3537 to vary, making some of the tests fail. I have changed pcretest so that:
3539 (a) When it is outputting text in the compiled version of a pattern, bytes
3540 other than 32-126 are always shown as hex escapes.
3542 (b) When it is outputting text that is a matched part of a subject string,
3543 it does the same, unless a different locale has been set for the match
3544 (using the /L modifier). In this case, it uses isprint() to decide.
3546 4. Fixed a major bug that caused incorrect computation of the amount of memory
3547 required for a compiled pattern when options that changed within the
3548 pattern affected the logic of the preliminary scan that determines the
3549 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3550 that the computed length was too small. The symptoms of this bug were
3551 either the PCRE error "internal error: code overflow" from pcre_compile(),
3552 or a glibc crash with a message such as "pcretest: free(): invalid next
3553 size (fast)". Examples of patterns that provoked this bug (shown in
3554 pcretest format) are:
3556 /(?-x: )/x
3557 /(?x)(?-x: \s*#\s*)/
3558 /((?i)[\x{c0}])/8
3559 /(?i:[\x{c0}])/8
3561 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3562 is now done differently.
3564 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3565 wrapper classes; (b) implement a new function in the C++ scanner that is
3566 more efficient than the old way of doing things because it avoids levels of
3567 recursion in the regex matching; (c) add a paragraph to the documentation
3568 for the FullMatch() function.
3570 6. The escape sequence \n was being treated as whatever was defined as
3571 "newline". Not only was this contrary to the documentation, which states
3572 that \n is character 10 (hex 0A), but it also went horribly wrong when
3573 "newline" was defined as CRLF. This has been fixed.
3575 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3576 was being set to -1 for the "end of line" case (supposedly a value that no
3577 character can have). Though this value is never used (the check for end of
3578 line is "zero bytes in current character"), it caused compiler complaints.
3579 I've changed it to 0xffffffff.
3581 8. In pcre_version.c, the version string was being built by a sequence of
3582 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3583 string (as it is for production releases) called a macro with an empty
3584 argument. The C standard says the result of this is undefined. The gcc
3585 compiler treats it as an empty string (which was what was wanted) but it is
3586 reported that Visual C gives an error. The source has been hacked around to
3587 avoid this problem.
3589 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3590 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3591 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3592 of them did).
3594 10. Originally, pcretest opened its input and output without "b"; then I was
3595 told that "b" was needed in some environments, so it was added for release
3596 5.0 to both the input and output. (It makes no difference on Unix-like
3597 systems.) Later I was told that it is wrong for the input on Windows. I've
3598 now abstracted the modes into two macros, to make it easier to fiddle with
3599 them, and removed "b" from the input mode under Windows.
3601 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3603 12. Added -help and --help to pcretest as an official way of being reminded
3604 of the options.
3606 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3607 and pcrecpp.cc because they annoy compilers at high warning levels.
3609 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3611 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3612 scripts are not C programs :-) and which was not noticed because it works
3613 on Linux.
3615 16. pcretest is supposed to handle any length of pattern and data line (as one
3616 line or as a continued sequence of lines) by extending its input buffer if
3617 necessary. This feature was broken for very long pattern lines, leading to
3618 a string of junk being passed to pcre_compile() if the pattern was longer
3619 than about 50K.
3621 17. I have done a major re-factoring of the way pcre_compile() computes the
3622 amount of memory needed for a compiled pattern. Previously, there was code
3623 that made a preliminary scan of the pattern in order to do this. That was
3624 OK when PCRE was new, but as the facilities have expanded, it has become
3625 harder and harder to keep it in step with the real compile phase, and there
3626 have been a number of bugs (see for example, 4 above). I have now found a
3627 cunning way of running the real compile function in a "fake" mode that
3628 enables it to compute how much memory it would need, while actually only
3629 ever using a few hundred bytes of working memory and without too many
3630 tests of the mode. This should make future maintenance and development
3631 easier. A side effect of this work is that the limit of 200 on the nesting
3632 depth of parentheses has been removed (though this was never a serious
3633 limitation, I suspect). However, there is a downside: pcre_compile() now
3634 runs more slowly than before (30% or more, depending on the pattern). I
3635 hope this isn't a big issue. There is no effect on runtime performance.
3637 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3638 newline (only possible for the last line of a file) and it was a
3639 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3641 19. Added additional timing features to pcretest. (1) The -tm option now times
3642 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3643 separate command line item, by a number that specifies the number of
3644 repeats to use when timing. The default is 50000; this gives better
3645 precision, but takes uncomfortably long for very large patterns.
3647 20. Extended pcre_study() to be more clever in cases where a branch of a
3648 subpattern has no definite first character. For example, (a*|b*)[cd] would
3649 previously give no result from pcre_study(). Now it recognizes that the
3650 first character must be a, b, c, or d.
3652 21. There was an incorrect error "recursive call could loop indefinitely" if
3653 a subpattern (or the entire pattern) that was being tested for matching an
3654 empty string contained only one non-empty item after a nested subpattern.
3655 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3656 incorrectly, because the \d was being skipped in the check.
3658 22. The pcretest program now has a new pattern option /B and a command line
3659 option -b, which is equivalent to adding /B to every pattern. This causes
3660 it to show the compiled bytecode, without the additional information that
3661 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3662 is the same as /B/I).
3664 23. A new optimization is now able automatically to treat some sequences such
3665 as a*b as a*+b. More specifically, if something simple (such as a character
3666 or a simple class like \d) has an unlimited quantifier, and is followed by
3667 something that cannot possibly match the quantified thing, the quantifier
3668 is automatically "possessified".
3670 24. A recursive reference to a subpattern whose number was greater than 39
3671 went wrong under certain circumstances in UTF-8 mode. This bug could also
3672 have affected the operation of pcre_study().
3674 25. Realized that a little bit of performance could be had by replacing
3675 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3677 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3679 27. Possessive quantifiers such as a++ were previously implemented by turning
3680 them into atomic groups such as ($>a+). Now they have their own opcodes,
3681 which improves performance. This includes the automatically created ones
3682 from 23 above.
3684 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3685 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3686 the first matched character to be a colon. This applied both to named and
3687 numbered groups.
3689 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3691 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3692 building PCRE on Windows easier, so I have included it in the distribution.
3694 31. There is now a check in pcretest against a ridiculously large number being
3695 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3696 loop, the loop is abandoned.
3698 32. Forward references to subpatterns in conditions such as (?(2)...) where
3699 subpattern 2 is defined later cause pcre_compile() to search forwards in
3700 the pattern for the relevant set of parentheses. This search went wrong
3701 when there were unescaped parentheses in a character class, parentheses
3702 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3704 33. "Subroutine" calls and backreferences were previously restricted to
3705 referencing subpatterns earlier in the regex. This restriction has now
3706 been removed.
3708 34. Added a number of extra features that are going to be in Perl 5.10. On the
3709 whole, these are just syntactic alternatives for features that PCRE had
3710 previously implemented using the Python syntax or my own invention. The
3711 other formats are all retained for compatibility.
3713 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3714 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3715 also .NET compatible.
3717 (b) A recursion or subroutine call to a named group can now be defined as
3718 (?&name) as well as (?P>name).
3720 (c) A backreference to a named group can now be defined as \k<name> or
3721 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3722 5.10, are also .NET compatible.
3724 (d) A conditional reference to a named group can now use the syntax
3725 (?(<name>) or (?('name') as well as (?(name).
3727 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3728 groups (named and numbered) that are never evaluated inline, but can be
3729 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3730 is always false. There may be only one alternative in such a group.
3732 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3733 as the simple (?(R). The condition is true only if the most recent
3734 recursion is that of the given number or name. It does not search out
3735 through the entire recursion stack.
3737 (g) The escape \gN or \g{N} has been added, where N is a positive or
3738 negative number, specifying an absolute or relative reference.
3740 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3741 some "unreachable code" warnings.
3743 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3744 things, this adds five new scripts.
3746 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3747 There were also incompatibilities regarding the handling of \Q..\E inside
3748 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3749 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3751 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3752 matches an empty string, and forcibly breaks the loop. There were bugs in
3753 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3754 against aaaa the result was just "a" rather than "aaaa", for example. Two
3755 separate and independent bugs (that affected different cases) have been
3756 fixed.
3758 39. Refactored the code to abolish the use of different opcodes for small
3759 capturing bracket numbers. This is a tidy that I avoided doing when I
3760 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3761 The new approach is not only tidier, it makes it possible to reduce the
3762 memory needed to fix the previous bug (38).
3764 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3765 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3766 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3767 mode.
3769 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3770 report.
3772 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3773 copy construction and assignment in the C++ wrapper.
3775 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3776 couple of bugs that could have given wrong results in the "--newline=crlf"
3777 case.
3779 44. Added a number of casts and did some reorganization of signed/unsigned int
3780 variables following suggestions from Dair Grant. Also renamed the variable
3781 "this" as "item" because it is a C++ keyword.
3783 45. Arranged for dftables to add
3785 #include "pcre_internal.h"
3787 to pcre_chartables.c because without it, gcc 4.x may remove the array
3788 definition from the final binary if PCRE is built into a static library and
3789 dead code stripping is activated.
3791 46. For an unanchored pattern, if a match attempt fails at the start of a
3792 newline sequence, and the newline setting is CRLF or ANY, and the next two
3793 characters are CRLF, advance by two characters instead of one.
3796 Version 6.7 04-Jul-06
3797 ---------------------
3799 1. In order to handle tests when input lines are enormously long, pcretest has
3800 been re-factored so that it automatically extends its buffers when
3801 necessary. The code is crude, but this _is_ just a test program. The
3802 default size has been increased from 32K to 50K.
3804 2. The code in pcre_study() was using the value of the re argument before
3805 testing it for NULL. (Of course, in any sensible call of the function, it
3806 won't be NULL.)
3808 3. The memmove() emulation function in pcre_internal.h, which is used on
3809 systems that lack both memmove() and bcopy() - that is, hardly ever -
3810 was missing a "static" storage class specifier.
3812 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3813 containing an extended class (one that cannot be represented by a bitmap
3814 because it contains high-valued characters or Unicode property items, e.g.
3815 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3816 pattern, but PCRE should not loop if you do not (it no longer does).
3817 [Detail: two cases were found: (a) a repeated subpattern containing an
3818 extended class; (b) a recursive reference to a subpattern that followed a
3819 previous extended class. It wasn't skipping over the extended class
3820 correctly when UTF-8 mode was not set.]
3822 5. A negated single-character class was not being recognized as fixed-length
3823 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3824 compile error "lookbehind assertion is not fixed length".
3826 6. The RunPerlTest auxiliary script was showing an unexpected difference
3827 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3828 write a Perl script that can interpret lines of an input file either as
3829 byte characters or as UTF-8, which is what "perltest" was being required to
3830 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3831 can't do is switch easily at run time between having the "use utf8;" pragma
3832 or not. In the end, I fudged it by using the RunPerlTest script to insert
3833 "use utf8;" explicitly for the UTF-8 tests.
3835 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3836 the end of the subject string, contrary to the documentation and to what
3837 Perl does. This was true of both matching functions. Now it matches only at
3838 the start of the subject and immediately after *internal* newlines.
3840 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3841 a pointer to an int instead of a pointer to an unsigned long int. This
3842 caused problems on 64-bit systems.
3844 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3845 instance of the 'standard' template library not being so standard".
3847 10. There was no check on the number of named subpatterns nor the maximum
3848 length of a subpattern name. The product of these values is used to compute
3849 the size of the memory block for a compiled pattern. By supplying a very
3850 long subpattern name and a large number of named subpatterns, the size
3851 computation could be caused to overflow. This is now prevented by limiting
3852 the length of names to 32 characters, and the number of named subpatterns
3853 to 10,000.
3855 11. Subpatterns that are repeated with specific counts have to be replicated in
3856 the compiled pattern. The size of memory for this was computed from the
3857 length of the subpattern and the repeat count. The latter is limited to
3858 65535, but there was no limit on the former, meaning that integer overflow
3859 could in principle occur. The compiled length of a repeated subpattern is
3860 now limited to 30,000 bytes in order to prevent this.
3862 12. Added the optional facility to have named substrings with the same name.
3864 13. Added the ability to use a named substring as a condition, using the
3865 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3866 are numbers (not recommended). Forward references are permitted.
3868 14. Added forward references in named backreferences (if you see what I mean).
3870 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3871 pattern could run off the end of the subject. For example, the pattern
3872 "(?s)(.{1,5})"8 did this with the subject "ab".
3874 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3875 PCRE_CASELESS was set when matching characters that were quantified with ?
3876 or *.
3878 17. A character class other than a single negated character that had a minimum
3879 but no maximum quantifier - for example [ab]{6,} - was not handled
3880 correctly by pce_dfa_exec(). It would match only one character.
3882 18. A valid (though odd) pattern that looked like a POSIX character
3883 class but used an invalid character after [ (for example [[,abc,]]) caused
3884 pcre_compile() to give the error "Failed: internal error: code overflow" or
3885 in some cases to crash with a glibc free() error. This could even happen if
3886 the pattern terminated after [[ but there just happened to be a sequence of
3887 letters, a binary zero, and a closing ] in the memory that followed.
3889 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3890 over the years. Originally (before any Unicode support), just the bottom 8
3891 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3892 output from "man perlunicode" includes this:
3894 The regular expression compiler produces polymorphic opcodes. That
3895 is, the pattern adapts to the data and automatically switches to
3896 the Unicode character scheme when presented with Unicode data--or
3897 instead uses a traditional byte scheme when presented with byte
3898 data.
3900 Sadly, a wide octal escape does not cause a switch, and in a string with
3901 no other multibyte characters, these octal escapes are treated as before.
3902 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3903 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3904 Unicode string.
3906 I have not perpetrated such confusion in PCRE. Up till now, it took just
3907 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3908 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3909 translate to the appropriate multibyte character.
3911 29. Applied some refactoring to reduce the number of warnings from Microsoft
3912 and Borland compilers. This has included removing the fudge introduced
3913 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3914 a warning about an unused variable.
3916 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3917 characters since release 4.0, because Perl (from release 5.004) does not.
3918 [Or at least, is documented not to: some releases seem to be in conflict
3919 with the documentation.] However, when a pattern was studied with
3920 pcre_study() and all its branches started with \s, PCRE still included VT
3921 as a possible starting character. Of course, this did no harm; it just
3922 caused an unnecessary match attempt.
3924 22. Removed a now-redundant internal flag bit that recorded the fact that case
3925 dependency changed within the pattern. This was once needed for "required
3926 byte" processing, but is no longer used. This recovers a now-scarce options
3927 bit. Also moved the least significant internal flag bit to the most-
3928 significant bit of the word, which was not previously used (hangover from
3929 the days when it was an int rather than a uint) to free up another bit for
3930 the future.
3932 23. Added support for CRLF line endings as well as CR and LF. As well as the
3933 default being selectable at build time, it can now be changed at runtime
3934 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3935 specify that it is scanning data with non-default line endings.
3937 24. Changed the definition of CXXLINK to make it agree with the definition of
3938 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3940 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3941 recursions. This makes a big different to stack usage for some patterns.
3943 26. If a subpattern containing a named recursion or subroutine reference such
3944 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3945 the space required for the compiled pattern went wrong and gave too small a
3946 value. Depending on the environment, this could lead to "Failed: internal
3947 error: code overflow at offset 49" or "glibc detected double free or
3948 corruption" errors.
3950 27. Applied patches from Google (a) to support the new newline modes and (b) to
3951 advance over multibyte UTF-8 characters in GlobalReplace.
3953 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3954 difference for some implementation of PCRE in some Windows version.
3956 29. Added some extra testing facilities to pcretest:
3958 \q<number> in a data line sets the "match limit" value
3959 \Q<number> in a data line sets the "match recursion limt" value
3960 -S <number> sets the stack size, where <number> is in megabytes
3962 The -S option isn't available for Windows.
3965 Version 6.6 06-Feb-06
3966 ---------------------
3968 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3969 in pcreposix.h. I have copied the definition from pcre.h.
3971 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3972 because pcre.h is no longer a built file.
3974 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3975 not normally included in the compiled code.
3978 Version 6.5 01-Feb-06
3979 ---------------------
3981 1. When using the partial match feature with pcre_dfa_exec(), it was not
3982 anchoring the second and subsequent partial matches at the new starting
3983 point. This could lead to incorrect results. For example, with the pattern
3984 /1234/, partially matching against "123" and then "a4" gave a match.
3986 2. Changes to pcregrep:
3988 (a) All non-match returns from pcre_exec() were being treated as failures
3989 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3990 error message is output. Some extra information is given for the
3992 probably the only errors that are likely to be caused by users (by
3993 specifying a regex that has nested indefinite repeats, for instance).
3994 If there are more than 20 of these errors, pcregrep is abandoned.
3996 (b) A binary zero was treated as data while matching, but terminated the
3997 output line if it was written out. This has been fixed: binary zeroes
3998 are now no different to any other data bytes.
4000 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
4001 used to set a locale for matching. The --locale=xxxx long option has
4002 been added (no short equivalent) to specify a locale explicitly on the
4003 pcregrep command, overriding the environment variables.
4005 (d) When -B was used with -n, some line numbers in the output were one less
4006 than they should have been.
4008 (e) Added the -o (--only-matching) option.
4010 (f) If -A or -C was used with -c (count only), some lines of context were
4011 accidentally printed for the final match.
4013 (g) Added the -H (--with-filename) option.
4015 (h) The combination of options -rh failed to suppress file names for files
4016 that were found from directory arguments.
4018 (i) Added the -D (--devices) and -d (--directories) options.
4020 (j) Added the -F (--fixed-strings) option.
4022 (k) Allow "-" to be used as a file name for -f as well as for a data file.
4024 (l) Added the --colo(u)r option.
4026 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
4027 is not present by default.
4029 3. A nasty bug was discovered in the handling of recursive patterns, that is,
4030 items such as (?R) or (?1), when the recursion could match a number of
4031 alternatives. If it matched one of the alternatives, but subsequently,
4032 outside the recursion, there was a failure, the code tried to back up into
4033 the recursion. However, because of the way PCRE is implemented, this is not
4034 possible, and the result was an incorrect result from the match.
4036 In order to prevent this happening, the specification of recursion has
4037 been changed so that all such subpatterns are automatically treated as
4038 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
4040 4. I had overlooked the fact that, in some locales, there are characters for
4041 which isalpha() is true but neither isupper() nor islower() are true. In
4042 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
4043 and ordfeminine) are like this. This affected the treatment of \w and \W
4044 when they appeared in character classes, but not when they appeared outside
4045 a character class. The bit map for "word" characters is now created
4046 separately from the results of isalnum() instead of just taking it from the
4047 upper, lower, and digit maps. (Plus the underscore character, of course.)
4049 5. The above bug also affected the handling of POSIX character classes such as
4050 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
4051 permanent tables. Instead, the bit maps for such a class were previously
4052 created as the appropriate unions of the upper, lower, and digit bitmaps.
4053 Now they are created by subtraction from the [[:word:]] class, which has
4054 its own bitmap.
4056 6. The [[:blank:]] character class matches horizontal, but not vertical space.
4057 It is created by subtracting the vertical space characters (\x09, \x0a,
4058 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
4059 subtraction was done in the overall bitmap for a character class, meaning
4060 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
4061 be recognized. This bug has been fixed.
4063 7. Patches from the folks at Google:
4065 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
4066 real life, but is still worth protecting against".
4068 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
4069 regular expressions".
4071 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
4072 have it.
4074 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
4075 "configure" and the latter not, in order to fix a problem somebody had
4076 with compiling the Arg class on HP-UX.
4078 (e) Improve the error-handling of the C++ wrapper a little bit.
4080 (f) New tests for checking recursion limiting.
4082 8. The pcre_memmove() function, which is used only if the environment does not
4083 have a standard memmove() function (and is therefore rarely compiled),
4084 contained two bugs: (a) use of int instead of size_t, and (b) it was not
4085 returning a result (though PCRE never actually uses the result).
4087 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
4088 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
4089 returned instead of calling malloc() with an overflowing number that would
4090 most likely cause subsequent chaos.
4092 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
4094 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
4095 with this option is matched, the nmatch and pmatch options of regexec() are
4096 ignored.
4098 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
4099 provided in case anyone wants to the the POSIX interface with UTF-8
4100 strings.
4102 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
4103 C++ linking (needed for some HP-UX environments).
4105 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
4106 (unused parameter) and in the pcre_printint() function (omitted "default"
4107 switch label when the default is to do nothing).
4109 15. Added some code to make it possible, when PCRE is compiled as a C++
4110 library, to replace subject pointers for pcre_exec() with a smart pointer
4111 class, thus making it possible to process discontinuous strings.
4113 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
4114 much the same function. They were added by different people who were trying
4115 to make PCRE easy to compile on non-Unix systems. It has been suggested
4116 that PCRE_EXPORT be abolished now that there is more automatic apparatus
4117 for compiling on Windows systems. I have therefore replaced it with
4118 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
4119 defaults to "extern" for C or "extern C" for C++, which works fine on
4120 Unix-like systems. It is now possible to override the value of PCRE_DATA_
4121 SCOPE with something explicit in config.h. In addition:
4123 (a) pcreposix.h still had just "extern" instead of either of these macros;
4124 I have replaced it with PCRE_DATA_SCOPE.
4126 (b) Functions such as _pcre_xclass(), which are internal to the library,
4127 but external in the C sense, all had PCRE_EXPORT in their definitions.
4128 This is apparently wrong for the Windows case, so I have removed it.
4129 (It makes no difference on Unix-like systems.)
4131 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
4132 of recursive calls to match(). This is different to MATCH_LIMIT because
4133 that limits the total number of calls to match(), not all of which increase
4134 the depth of recursion. Limiting the recursion depth limits the amount of
4135 stack (or heap if NO_RECURSE is set) that is used. The default can be set
4136 when PCRE is compiled, and changed at run time. A patch from Google adds
4137 this functionality to the C++ interface.
4139 18. Changes to the handling of Unicode character properties:
4141 (a) Updated the table to Unicode 4.1.0.
4143 (b) Recognize characters that are not in the table as "Cn" (undefined).
4145 (c) I revised the way the table is implemented to a much improved format
4146 which includes recognition of ranges. It now supports the ranges that
4147 are defined in UnicodeData.txt, and it also amalgamates other
4148 characters into ranges. This has reduced the number of entries in the
4149 table from around 16,000 to around 3,000, thus reducing its size
4150 considerably. I realized I did not need to use a tree structure after
4151 all - a binary chop search is just as efficient. Having reduced the
4152 number of entries, I extended their size from 6 bytes to 8 bytes to
4153 allow for more data.
4155 (d) Added support for Unicode script names via properties such as \p{Han}.
4157 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4158 matching that character.
4160 20. When matching a repeated Unicode property with a minimum greater than zero,
4161 (for example \pL{2,}), PCRE could look past the end of the subject if it
4162 reached it while seeking the minimum number of characters. This could
4163 happen only if some of the characters were more than one byte long, because
4164 there is a check for at least the minimum number of bytes.
4166 21. Refactored the implementation of \p and \P so as to be more general, to
4167 allow for more different types of property in future. This has changed the
4168 compiled form incompatibly. Anybody with saved compiled patterns that use
4169 \p or \P will have to recompile them.
4171 22. Added "Any" and "L&" to the supported property types.
4173 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4174 but give a compile time error if the value is greater than 0xff.
4176 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4177 accidentally not being installed or uninstalled.
4179 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4180 made were to insert the current release number. This seemed silly, because
4181 it made things harder for people building PCRE on systems that don't run
4182 "configure". I have turned pcre.h into a distributed file, no longer built
4183 by "configure", with the version identification directly included. There is
4184 no longer a pcre.h.in file.
4186 However, this change necessitated a change to the pcre-config script as
4187 well. It is built from pcre-config.in, and one of the substitutions was the
4188 release number. I have updated configure.ac so that ./configure now finds
4189 the release number by grepping pcre.h.
4191 26. Added the ability to run the tests under valgrind.
4194 Version 6.4 05-Sep-05
4195 ---------------------
4197 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4198 "--" to be printed when multiple files were scanned, even when none of the
4199 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4200 consider it to be a bug, and have restored the previous behaviour.
4202 2. A couple of code tidies to get rid of compiler warnings.
4204 3. The pcretest program used to cheat by referring to symbols in the library
4205 whose names begin with _pcre_. These are internal symbols that are not
4206 really supposed to be visible externally, and in some environments it is
4207 possible to suppress them. The cheating is now confined to including
4208 certain files from the library's source, which is a bit cleaner.
4210 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4211 file's purpose clearer.
4213 5. Reorganized pcre_ucp_findchar().
4216 Version 6.3 15-Aug-05
4217 ---------------------
4219 1. The file libpcre.pc.in did not have general read permission in the tarball.
4221 2. There were some problems when building without C++ support:
4223 (a) If C++ support was not built, "make install" and "make test" still
4224 tried to test it.
4226 (b) There were problems when the value of CXX was explicitly set. Some
4227 changes have been made to try to fix these, and ...
4229 (c) --disable-cpp can now be used to explicitly disable C++ support.
4231 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4232 backslash in a target when C++ was disabled. This confuses some
4233 versions of "make", apparently. Using an intermediate variable solves
4234 this. (Same for CPP_LOBJ.)
4236 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4237 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4238 necessary on certain architectures.
4240 4. Added a setting of -export-symbols-regex to the link command to remove
4241 those symbols that are exported in the C sense, but actually are local
4242 within the library, and not documented. Their names all begin with
4243 "_pcre_". This is not a perfect job, because (a) we have to except some
4244 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4245 available (and never for static libraries). I have made a note to try to
4246 find a way round (a) in the future.
4249 Version 6.2 01-Aug-05
4250 ---------------------
4252 1. There was no test for integer overflow of quantifier values. A construction
4253 such as {1111111111111111} would give undefined results. What is worse, if
4254 a minimum quantifier for a parenthesized subpattern overflowed and became
4255 negative, the calculation of the memory size went wrong. This could have
4256 led to memory overwriting.
4258 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4260 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4261 operating environments where this matters.
4263 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4264 PCRE options from within the C++ wrapper.
4266 5. Named capturing subpatterns were not being correctly counted when a pattern
4267 was compiled. This caused two problems: (a) If there were more than 100
4268 such subpatterns, the calculation of the memory needed for the whole
4269 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4270 back references of the form \12, where the number was greater than 9, were
4271 not recognized as back references, even though there were sufficient
4272 previous subpatterns.
4274 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4275 versions of gcc, e.g. 2.95.4.
4278 Version 6.1 21-Jun-05
4279 ---------------------
4281 1. There was one reference to the variable "posix" in pcretest.c that was not
4282 surrounded by "#if !defined NOPOSIX".
4284 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4285 the cross-check on the old pcre_info() function, for the benefit of the
4286 cut-down version of PCRE that is currently imported into Exim.
4288 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4289 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4290 patterns that start (?i) but I don't think that matters. The use of (?i) is
4291 just an example; this all applies to the other options as well.
4293 4. Since libtool seems to echo the compile commands it is issuing, the output
4294 from "make" can be reduced a bit by putting "@" in front of each libtool
4295 compile command.
4297 5. Patch from the folks at Google for configure.in to be a bit more thorough
4298 in checking for a suitable C++ installation before trying to compile the
4299 C++ stuff. This should fix a reported problem when a compiler was present,
4300 but no suitable headers.
4302 6. The man pages all had just "PCRE" as their title. I have changed them to
4303 be the relevant file name. I have also arranged that these names are
4304 retained in the file doc/pcre.txt, which is a concatenation in text format
4305 of all the man pages except the little individual ones for each function.
4307 7. The NON-UNIX-USE file had not been updated for the different set of source
4308 files that come with release 6. I also added a few comments about the C++
4309 wrapper.
4312 Version 6.0 07-Jun-05
4313 ---------------------
4315 1. Some minor internal re-organization to help with my DFA experiments.
4317 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4318 didn't matter for the library itself when fully configured, but did matter
4319 when compiling without UCP support, or within Exim, where the ucp files are
4320 not imported.
4322 3. Refactoring of the library code to split up the various functions into
4323 different source modules. The addition of the new DFA matching code (see
4324 below) to a single monolithic source would have made it really too
4325 unwieldy, quite apart from causing all the code to be include in a
4326 statically linked application, when only some functions are used. This is
4327 relevant even without the DFA addition now that patterns can be compiled in
4328 one application and matched in another.
4330 The downside of splitting up is that there have to be some external
4331 functions and data tables that are used internally in different modules of
4332 the library but which are not part of the API. These have all had their
4333 names changed to start with "_pcre_" so that they are unlikely to clash
4334 with other external names.
4336 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4337 a different (DFA) algorithm. Although it is slower than the original
4338 function, it does have some advantages for certain types of matching
4339 problem.
4341 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4342 including restarting after a partial match.
4344 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4345 defined when compiling for Windows was sent to me. I have put it into the
4346 code, though I have no means of testing or verifying it.
4348 7. Added the pcre_refcount() auxiliary function.
4350 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4351 match before or at the first newline in the subject string. In pcretest,
4352 the /f option on a pattern can be used to set this.
4354 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4355 would behave wrongly. This has been present in PCRE since release 4.0.
4357 10. A number of changes to the pcregrep command:
4359 (a) Refactored how -x works; insert ^(...)$ instead of setting
4360 PCRE_ANCHORED and checking the length, in preparation for adding
4361 something similar for -w.
4363 (b) Added the -w (match as a word) option.
4365 (c) Refactored the way lines are read and buffered so as to have more
4366 than one at a time available.
4368 (d) Implemented a pcregrep test script.
4370 (e) Added the -M (multiline match) option. This allows patterns to match
4371 over several lines of the subject. The buffering ensures that at least
4372 8K, or the rest of the document (whichever is the shorter) is available
4373 for matching (and similarly the previous 8K for lookbehind assertions).
4375 (f) Changed the --help output so that it now says
4377 -w, --word-regex(p)
4379 instead of two lines, one with "regex" and the other with "regexp"
4380 because that confused at least one person since the short forms are the
4381 same. (This required a bit of code, as the output is generated
4382 automatically from a table. It wasn't just a text change.)
4384 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4385 option but starts with a hyphen. Could be a pattern or a path name
4386 starting with a hyphen, for instance.
4388 (h) "-" can be given as a file name to represent stdin.
4390 (i) When file names are being printed, "(standard input)" is used for
4391 the standard input, for compatibility with GNU grep. Previously
4392 "<stdin>" was used.
4394 (j) The option --label=xxx can be used to supply a name to be used for
4395 stdin when file names are being printed. There is no short form.
4397 (k) Re-factored the options decoding logic because we are going to add
4398 two more options that take data. Such options can now be given in four
4399 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4401 (l) Added the -A, -B, and -C options for requesting that lines of context
4402 around matches be printed.
4404 (m) Added the -L option to print the names of files that do not contain
4405 any matching lines, that is, the complement of -l.
4407 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4408 continue to scan other files.
4410 (o) The -s option was incorrectly implemented. For compatibility with other
4411 greps, it now suppresses the error message for a non-existent or non-
4412 accessible file (but not the return code). There is a new option called
4413 -q that suppresses the output of matching lines, which was what -s was
4414 previously doing.
4416 (p) Added --include and --exclude options to specify files for inclusion
4417 and exclusion when recursing.
4419 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4420 Hopefully, it now does.
4422 12. Missing cast in pcre_study().
4424 13. Added an "uninstall" target to the makefile.
4426 14. Replaced "extern" in the function prototypes in Makefile.in with
4427 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4428 world, but is set differently for Windows.
4430 15. Added a second compiling function called pcre_compile2(). The only
4431 difference is that it has an extra argument, which is a pointer to an
4432 integer error code. When there is a compile-time failure, this is set
4433 non-zero, in addition to the error test pointer being set to point to an
4434 error message. The new argument may be NULL if no error number is required
4435 (but then you may as well call pcre_compile(), which is now just a
4436 wrapper). This facility is provided because some applications need a
4437 numeric error indication, but it has also enabled me to tidy up the way
4438 compile-time errors are handled in the POSIX wrapper.
4440 16. Added VPATH=.libs to the makefile; this should help when building with one
4441 prefix path and installing with another. (Or so I'm told by someone who
4442 knows more about this stuff than I do.)
4444 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4445 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4446 match everything, including newlines. This is not POSIX-compatible, but
4447 somebody wanted the feature. From pcretest it can be activated by using
4448 both the P and the s flags.
4450 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4452 19. libpcre.pc was being incorrectly installed as executable.
4454 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4455 it now also looks for '\r' so that it will work unmodified on Windows.
4457 21. Added Google's contributed C++ wrapper to the distribution.
4459 22. Added some untidy missing memory free() calls in pcretest, to keep
4460 Electric Fence happy when testing.
4464 Version 5.0 13-Sep-04
4465 ---------------------
4467 1. Internal change: literal characters are no longer packed up into items
4468 containing multiple characters in a single byte-string. Each character
4469 is now matched using a separate opcode. However, there may be more than one
4470 byte in the character in UTF-8 mode.
4472 2. The pcre_callout_block structure has two new fields: pattern_position and
4473 next_item_length. These contain the offset in the pattern to the next match
4474 item, and its length, respectively.
4476 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4477 insertion of callouts before each pattern item. Added the /C option to
4478 pcretest to make use of this.
4480 4. On the advice of a Windows user, the lines
4482 #if defined(_WIN32) || defined(WIN32)
4483 _setmode( _fileno( stdout ), 0x8000 );
4484 #endif /* defined(_WIN32) || defined(WIN32) */
4486 have been added to the source of pcretest. This apparently does useful
4487 magic in relation to line terminators.
4489 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4490 for the benefit of those environments where the "b" makes a difference.
4492 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4493 to know about it. I have put a hack into configure.in that adds in code
4494 to set GCC=yes if CC=icc. This seems to end up at a point in the
4495 generated configure script that is early enough to affect the setting of
4496 compiler options, which is what is needed, but I have no means of testing
4497 whether it really works. (The user who reported this had patched the
4498 generated configure script, which of course I cannot do.)
4500 LATER: After change 22 below (new libtool files), the configure script
4501 seems to know about icc (and also ecc). Therefore, I have commented out
4502 this hack in configure.in.
4504 7. Added support for pkg-config (2 patches were sent in).
4506 8. Negated POSIX character classes that used a combination of internal tables
4507 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4508 [[:^ascii]]. Typically, they would match almost any characters. The other
4509 POSIX classes were not broken in this way.
4511 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4512 to find the match, as PCRE was deluded into thinking that the match had to
4513 start at the start point or following a newline. The same bug applied to
4514 patterns with negative forward assertions or any backward assertions
4515 preceding ".*" at the start, unless the pattern required a fixed first
4516 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4518 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4519 starting at the last subject character, bytes beyond the end of the subject
4520 string were read.
4522 11. Renamed the variable "class" as "classbits" to make life easier for C++
4523 users. (Previously there was a macro definition, but it apparently wasn't
4524 enough.)
4526 12. Added the new field "tables" to the extra data so that tables can be passed
4527 in at exec time, or the internal tables can be re-selected. This allows
4528 a compiled regex to be saved and re-used at a later time by a different
4529 program that might have everything at different addresses.
4531 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4532 -R library as well as a -L library.
4534 14. The debugging options of pcretest (-d on the command line or D on a
4535 pattern) showed incorrect output for anything following an extended class
4536 that contained multibyte characters and which was followed by a quantifier.
4538 15. Added optional support for general category Unicode character properties
4539 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4540 support. It adds about 90K to the size of the library. The meanings of the
4541 inbuilt class escapes such as \d and \s have NOT been changed.
4543 16. Updated pcredemo.c to include calls to free() to release the memory for the
4544 compiled pattern.
4546 17. The generated file chartables.c was being created in the source directory
4547 instead of in the building directory. This caused the build to fail if the
4548 source directory was different from the building directory, and was
4549 read-only.
4551 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4552 file. No doubt somebody will tell me if they don't make sense... Also added
4553 Dan Mooney's comments about building on OpenVMS.
4555 19. Added support for partial matching via the PCRE_PARTIAL option for
4556 pcre_exec() and the \P data escape in pcretest.
4558 20. Extended pcretest with 3 new pattern features:
4560 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4561 write the compiled pattern to the file whose name is "rest-of-line".
4562 This is a straight binary dump of the data, with the saved pointer to
4563 the character tables forced to be NULL. The study data, if any, is
4564 written too. After writing, pcretest reads a new pattern.
4566 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4567 compiled pattern from the given file. There must not be any
4568 occurrences of "<" in the file name (pretty unlikely); if there are,
4569 pcretest will instead treat the initial "<" as a pattern delimiter.
4570 After reading in the pattern, pcretest goes on to read data lines as
4571 usual.
4573 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4574 and 16-bit fields in a compiled pattern, to simulate a pattern that
4575 was compiled on a host of opposite endianness.
4577 21. The pcre-exec() function can now cope with patterns that were compiled on
4578 hosts of opposite endianness, with this restriction:
4580 As for any compiled expression that is saved and used later, the tables
4581 pointer field cannot be preserved; the extra_data field in the arguments
4582 to pcre_exec() should be used to pass in a tables address if a value
4583 other than the default internal tables were used at compile time.
4585 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4586 now diagnosed as an error. Previously, most of the time, a negative number
4587 would have been treated as zero, but if in addition "ovector" was passed as
4588 NULL, a crash could occur.
4590 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4591 new versions from the libtool 1.5 distribution (the last one is a copy of
4592 a file called libtool.m4). This seems to have fixed the need to patch
4593 "configure" to support Darwin 1.3 (which I used to do). However, I still
4594 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4595 workstation).
4597 24. Changed the PCRE licence to be the more standard "BSD" licence.
4600 Version 4.5 01-Dec-03
4601 ---------------------
4603 1. There has been some re-arrangement of the code for the match() function so
4604 that it can be compiled in a version that does not call itself recursively.
4605 Instead, it keeps those local variables that need separate instances for
4606 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4607 needs to "recurse". Keeping track of where control must go is done by means
4608 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4609 hide most of the details from the main code, and operates only if
4610 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4611 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4612 operating.
4614 To make it easier for callers to provide specially tailored get/free
4615 functions for this usage, two new functions, pcre_stack_malloc, and
4616 pcre_stack_free, are used. They are always called in strict stacking order,
4617 and the size of block requested is always the same.
4619 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4620 PCRE has been compiled to use the stack or the heap for recursion. The
4621 -C option of pcretest uses this to show which version is compiled.
4623 A new data escape \S, is added to pcretest; it causes the amounts of store
4624 obtained and freed by both kinds of malloc/free at match time to be added
4625 to the output.
4627 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4628 what's available on my current Linux desktop machine.
4630 3. When matching a UTF-8 string, the test for a valid string at the start has
4631 been extended. If start_offset is not zero, PCRE now checks that it points
4632 to a byte that is the start of a UTF-8 character. If not, it returns
4633 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4634 this is necessary because there may be backward assertions in the pattern.
4635 When matching the same subject several times, it may save resources to use
4636 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4638 4. The code for checking the validity of UTF-8 strings has been tightened so
4639 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4640 containing "overlong sequences".
4642 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4643 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4644 should have been "&", but it just so happened that all the cases this let
4645 through by mistake were picked up later in the function.
4647 6. I had used a variable called "isblank" - this is a C99 function, causing
4648 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4650 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4651 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4652 so the version is shown; (c) stop "make test" echoing "./RunTest".
4654 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4656 9. The prototype for memmove() for systems that don't have it was using
4657 size_t, but the inclusion of the header that defines size_t was later. I've
4658 moved the #includes for the C headers earlier to avoid this.
4660 10. Added some adjustments to the code to make it easier to compiler on certain
4661 special systems:
4663 (a) Some "const" qualifiers were missing.
4664 (b) Added the macro EXPORT before all exported functions; by default this
4665 is defined to be empty.
4666 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4667 that it reads its output file name as an argument instead of writing
4668 to the standard output and assuming this can be redirected.
4670 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4671 class containing characters with values greater than 255, PCRE compilation
4672 went into a loop.
4674 12. A recursive reference to a subpattern that was within another subpattern
4675 that had a minimum quantifier of zero caused PCRE to crash. For example,
4676 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4677 recursion. If the recursively-called subpattern itself had a zero repeat,
4678 that was OK.
4680 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4681 buffer into which it was copied (for escape processing) was still set at
4682 1024, so long lines caused crashes.
4684 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4685 "internal error: code overflow...". This applied to any character class
4686 that was followed by a possessive quantifier.
4688 15. Modified the Makefile to add libpcre.la as a prerequisite for
4689 libpcreposix.la because I was told this is needed for a parallel build to
4690 work.
4692 16. If a pattern that contained .* following optional items at the start was
4693 studied, the wrong optimizing data was generated, leading to matching
4694 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4695 matching string must start with a or b or c. The correct conclusion for
4696 this pattern is that a match can start with any character.
4699 Version 4.4 13-Aug-03
4700 ---------------------
4702 1. In UTF-8 mode, a character class containing characters with values between
4703 127 and 255 was not handled correctly if the compiled pattern was studied.
4704 In fixing this, I have also improved the studying algorithm for such
4705 classes (slightly).
4707 2. Three internal functions had redundant arguments passed to them. Removal
4708 might give a very teeny performance improvement.
4710 3. Documentation bug: the value of the capture_top field in a callout is *one
4711 more than* the number of the hightest numbered captured substring.
4713 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4714 in incorrectly linking with a previously installed version. They now link
4715 explicitly with libpcre.la.
4717 5. configure.in no longer needs to recognize Cygwin specially.
4719 6. A problem in pcre.in for Windows platforms is fixed.
4721 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4722 pcretest, it used to include the size of the study block as part of its
4723 output. Unfortunately, the structure contains a field that has a different
4724 size on different hardware architectures. This meant that the tests that
4725 showed this size failed. As the block is currently always of a fixed size,
4726 this information isn't actually particularly useful in pcretest output, so
4727 I have just removed it.
4729 8. Three pre-processor statements accidentally did not start in column 1.
4730 Sadly, there are *still* compilers around that complain, even though
4731 standard C has not required this for well over a decade. Sigh.
4733 9. In pcretest, the code for checking callouts passed small integers in the
4734 callout_data field, which is a void * field. However, some picky compilers
4735 complained about the casts involved for this on 64-bit systems. Now
4736 pcretest passes the address of the small integer instead, which should get
4737 rid of the warnings.
4739 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4740 both compile and run time, and gives an error if an invalid UTF-8 sequence
4741 is found. There is a option for disabling this check in cases where the
4742 string is known to be correct and/or the maximum performance is wanted.
4744 11. In response to a bug report, I changed one line in Makefile.in from
4746 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4747 to
4748 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4750 to look similar to other lines, but I have no way of telling whether this
4751 is the right thing to do, as I do not use Windows. No doubt I'll get told
4752 if it's wrong...
4755 Version 4.3 21-May-03
4756 ---------------------
4758 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4759 Makefile.
4761 2. Some refactoring to improve the quality of the code:
4763 (i) The utf8_table... variables are now declared "const".
4765 (ii) The code for \cx, which used the "case flipping" table to upper case
4766 lower case letters, now just substracts 32. This is ASCII-specific,
4767 but the whole concept of \cx is ASCII-specific, so it seems
4768 reasonable.
4770 (iii) PCRE was using its character types table to recognize decimal and
4771 hexadecimal digits in the pattern. This is silly, because it handles
4772 only 0-9, a-f, and A-F, but the character types table is locale-
4773 specific, which means strange things might happen. A private
4774 table is now used for this - though it costs 256 bytes, a table is
4775 much faster than multiple explicit tests. Of course, the standard
4776 character types table is still used for matching digits in subject
4777 strings against \d.
4779 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4780 ending in _t are). So I've renamed it as ESC_tee.
4782 3. The first argument for regexec() in the POSIX wrapper should have been
4783 defined as "const".
4785 4. Changed pcretest to use malloc() for its buffers so that they can be
4786 Electric Fenced for debugging.
4788 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4789 to read one or more bytes before the start of the subject string. Often this
4790 had no effect on PCRE's behaviour, but in some circumstances it could
4791 provoke a segmentation fault.
4793 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4794 to try to read one or more bytes before the start of the subject string.
4796 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4797 UTF-8 support could misbehave in various ways if the subject string
4798 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4799 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4800 back over UTF-8 characters.)
4803 Version 4.2 14-Apr-03
4804 ---------------------
4806 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4808 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4809 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4810 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4811 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4812 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4814 Note: automatic setting of the BUILD variables is not yet working
4815 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4816 compile-time but not at link-time
4817 [LINK]: use for linking executables only
4818 make different versions for Windows and non-Windows
4819 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4820 libraries
4821 [LINK_FOR_BUILD]: new variable
4822 [OBJEXT]: use throughout
4823 [EXEEXT]: use throughout
4824 <winshared>: new target
4825 <wininstall>: new target
4826 <dftables.o>: use native compiler
4827 <dftables>: use native linker
4828 <install>: handle Windows platform correctly
4829 <clean>: ditto
4830 <check>: ditto
4831 copy DLL to top builddir before testing
4833 As part of these changes, -no-undefined was removed again. This was reported
4834 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4835 in any case.
4837 3. Some tidies to get rid of compiler warnings:
4839 . In the match_data structure, match_limit was an unsigned long int, whereas
4840 match_call_count was an int. I've made them both unsigned long ints.
4842 . In pcretest the fact that a const uschar * doesn't automatically cast to
4843 a void * provoked a warning.
4845 . Turning on some more compiler warnings threw up some "shadow" variables
4846 and a few more missing casts.
4848 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4849 option, a class that contained a single character with a value between 128
4850 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4852 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4853 option, a class that contained several characters, but with at least one
4854 whose value was between 128 and 255 caused PCRE to crash.
4857 Version 4.1 12-Mar-03
4858 ---------------------
4860 1. Compiling with gcc -pedantic found a couple of places where casts were
4861 needed, and a string in dftables.c that was longer than standard compilers are
4862 required to support.
4864 2. Compiling with Sun's compiler found a few more places where the code could
4865 be tidied up in order to avoid warnings.
4867 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4868 first of these names is deprecated in the latest Autoconf in favour of the name
4869 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4870 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4871 analogy I have changed it to CFLAGS_FOR_BUILD.
4873 4. Added -no-undefined to the linking command in the Makefile, because this is
4874 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4875 linking step for the pcreposix library.
4877 5. PCRE was failing to diagnose the case of two named groups with the same
4878 name.
4880 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4881 literal character that is needed in the subject for a match, and scans along to
4882 ensure that it is present before embarking on the full matching process. This
4883 saves time in cases of nested unlimited repeats that are never going to match.
4884 Problem: the scan can take a lot of time if the subject is very long (e.g.
4885 megabytes), thus penalizing straightforward matches. It is now done only if the
4886 amount of subject to be scanned is less than 1000 bytes.
4888 7. A lesser problem with the same optimization is that it was recording the
4889 first character of an anchored pattern as "needed", thus provoking a search
4890 right along the subject, even when the first match of the pattern was going to
4891 fail. The "needed" character is now not set for anchored patterns, unless it
4892 follows something in the pattern that is of non-fixed length. Thus, it still
4893 fulfils its original purpose of finding quick non-matches in cases of nested
4894 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4897 Version 4.0 17-Feb-03
4898 ---------------------
4900 1. If a comment in an extended regex that started immediately after a meta-item
4901 extended to the end of string, PCRE compiled incorrect data. This could lead to
4902 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4904 2. Moved to autoconf 2.53 and libtool 1.4.2.
4906 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4907 the special perltest8 script is no longer needed - all the tests can be run
4908 from a single perltest script.
4910 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4911 by \s. It has now been removed in PCRE. This means it isn't recognized as
4912 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4913 class [:space:] *does* include VT, thereby creating a mess.
4915 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4916 space and tab.
4918 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4919 its new features into the main test script, reducing the number of scripts.
4921 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4922 were backward compatible, and made the (?i) apply to the whole pattern, as if
4923 /i were given. Now it behaves more logically, and applies the option setting
4924 only to what follows. PCRE has been changed to follow suit. However, if it
4925 finds options settings right at the start of the pattern, it extracts them into
4926 the global options, as before. Thus, they show up in the info data.
4928 8. Added support for the \Q...\E escape sequence. Characters in between are
4929 treated as literals. This is slightly different from Perl in that $ and @ are
4930 also handled as literals inside the quotes. In Perl, they will cause variable
4931 interpolation. Note the following examples:
4933 Pattern PCRE matches Perl matches
4935 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4936 \Qabc\$xyz\E abc\$xyz abc\$xyz
4937 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4939 For compatibility with Perl, \Q...\E sequences are recognized inside character
4940 classes as well as outside them.
4942 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4943 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4944 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4945 signed/unsigned warnings.
4947 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4948 option for pcretest, so I've replaced it by a simple function that does just
4949 that job.
4951 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4952 "pcregrep -".
4954 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4955 Java package. This provides some syntactic sugar for simple cases of what my
4956 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4957 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4958 item, you can use this simplified notation. Note that only makes sense with
4959 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4960 greediness, whatever the setting of the PCRE_UNGREEDY option.
4962 13. A change of greediness default within a pattern was not taking effect at
4963 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4964 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4965 was abstracted outside.
4967 14. PCRE now supports the \G assertion. It is true when the current matching
4968 position is at the start point of the match. This differs from \A when the
4969 starting offset is non-zero. Used with the /g option of pcretest (or similar
4970 code), it works in the same way as it does for Perl's /g option. If all
4971 alternatives of a regex begin with \G, the expression is anchored to the start
4972 match position, and the "anchored" flag is set in the compiled expression.
4974 15. Some bugs concerning the handling of certain option changes within patterns
4975 have been fixed. These applied to options other than (?ims). For example,
4976 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4977 been the other way round. Some of this was related to change 7 above.
4979 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4980 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4981 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4982 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4984 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4985 mode. Unlike ".", it always matches newline, whatever the setting of
4986 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4987 assertions. Perl allows it, but it doesn't (in general) work because it can't
4988 calculate the length of the lookbehind. At least, that's the case for Perl
4989 5.8.0 - I've been told they are going to document that it doesn't work in
4990 future.
4992 18. Added an error diagnosis for escapes that PCRE does not support: these are
4993 \L, \l, \N, \P, \p, \U, \u, and \X.
4995 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4996 reading past the end of the pattern in cases such as /[abcd/.
4998 20. PCRE was getting more memory than necessary for patterns with classes that
4999 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
5001 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
5002 compiling PCRE for use with Virtual Pascal.
5004 22. Small fix to the Makefile to make it work properly if the build is done
5005 outside the source tree.
5007 23. Added a new extension: a condition to go with recursion. If a conditional
5008 subpattern starts with (?(R) the "true" branch is used if recursion has
5009 happened, whereas the "false" branch is used only at the top level.
5011 24. When there was a very long string of literal characters (over 255 bytes
5012 without UTF support, over 250 bytes with UTF support), the computation of how
5013 much memory was required could be incorrect, leading to segfaults or other
5014 strange effects.
5016 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
5017 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
5018 there was a subsequent back reference to those brackets. This meant that, for
5019 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
5020 possible to check for precisely this case. All we can do is abandon the
5021 optimization if .* occurs inside capturing brackets when there are any back
5022 references whatsoever. (See below for a better fix that came later.)
5024 26. The handling of the optimization for finding the first character of a
5025 non-anchored pattern, and for finding a character that is required later in the
5026 match were failing in some cases. This didn't break the matching; it just
5027 failed to optimize when it could. The way this is done has been re-implemented.
5029 27. Fixed typo in error message for invalid (?R item (it said "(?p").
5031 28. Added a new feature that provides some of the functionality that Perl
5032 provides with (?{...}). The facility is termed a "callout". The way it is done
5033 in PCRE is for the caller to provide an optional function, by setting
5034 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
5035 global variable. By default it is unset, which disables all calling out. To get
5036 the function called, the regex must include (?C) at appropriate points. This
5037 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
5038 This provides a means of identifying different callout points. When PCRE
5039 reaches such a point in the regex, if pcre_callout has been set, the external
5040 function is called. It is provided with data in a structure called
5041 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
5042 matching continues; if it returns a non-zero value, the match at the current
5043 point fails. However, backtracking will occur if possible. [This was changed
5044 later and other features added - see item 49 below.]
5046 29. pcretest is upgraded to test the callout functionality. It provides a
5047 callout function that displays information. By default, it shows the start of
5048 the match and the current position in the text. There are some new data escapes
5049 to vary what happens:
5051 \C+ in addition, show current contents of captured substrings
5052 \C- do not supply a callout function
5053 \C!n return 1 when callout number n is reached
5054 \C!n!m return 1 when callout number n is reached for the mth time
5056 30. If pcregrep was called with the -l option and just a single file name, it
5057 output "<stdin>" if a match was found, instead of the file name.
5059 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
5060 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
5061 pcre_exec(). This saves a malloc/free per call. The default value of
5062 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
5063 when configuring.
5065 32. The default maximum size of a compiled pattern is 64K. There have been a
5066 few cases of people hitting this limit. The code now uses macros to handle the
5067 storing of links as offsets within the compiled pattern. It defaults to 2-byte
5068 links, but this can be changed to 3 or 4 bytes by --with-link-size when
5069 configuring. Tests 2 and 5 work only with 2-byte links because they output
5070 debugging information about compiled patterns.
5072 33. Internal code re-arrangements:
5074 (a) Moved the debugging function for printing out a compiled regex into
5075 its own source file (printint.c) and used #include to pull it into
5076 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
5077 separate copies.
5079 (b) Defined the list of op-code names for debugging as a macro in
5080 internal.h so that it is next to the definition of the opcodes.
5082 (c) Defined a table of op-code lengths for simpler skipping along compiled
5083 code. This is again a macro in internal.h so that it is next to the
5084 definition of the opcodes.
5086 34. Added support for recursive calls to individual subpatterns, along the
5087 lines of Robin Houston's patch (but implemented somewhat differently).
5089 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
5090 allow it to read and process whole directories in Win32. This code was
5091 contributed by Lionel Fourquaux; it has not been tested by me.
5093 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
5094 used to name a group. Names consist of alphanumerics and underscores, and must
5095 be unique. Back references use the syntax (?P=name) and recursive calls use
5096 (?P>name) which is a PCRE extension to the Python extension. Groups still have
5097 numbers. The function pcre_fullinfo() can be used after compilation to extract
5098 a name/number map. There are three relevant calls:
5100 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
5101 PCRE_INFO_NAMECOUNT yields the number of entries
5102 PCRE_INFO_NAMETABLE yields a pointer to the map.
5104 The map is a vector of fixed-size entries. The size of each entry depends on
5105 the length of the longest name used. The first two bytes of each entry are the
5106 group number, most significant byte first. There follows the corresponding
5107 name, zero terminated. The names are in alphabetical order.
5109 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
5110 case instead of 255. Making it the same both with and without UTF-8 support
5111 means that the same test output works with both.
5113 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
5114 calling malloc() with a zero argument.
5116 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
5117 optimization. I've improved things by keeping a bitmap of backreferences with
5118 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
5119 fact referenced, the optimization can be applied. It is unlikely that a
5120 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
5121 the match to follow \n) will appear inside brackets with a number greater than
5122 31, but if it does, any back reference > 31 suppresses the optimization.
5124 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
5125 of disabling numbered capturing parentheses. Any opening parenthesis that is
5126 not followed by ? behaves as if it were followed by ?: but named parentheses
5127 can still be used for capturing (and they will acquire numbers in the usual
5128 way).
5130 41. Redesigned the return codes from the match() function into yes/no/error so
5131 that errors can be passed back from deep inside the nested calls. A malloc
5132 failure while inside a recursive subpattern call now causes the
5133 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
5135 42. It is now possible to set a limit on the number of times the match()
5136 function is called in a call to pcre_exec(). This facility makes it possible to
5137 limit the amount of recursion and backtracking, though not in a directly
5138 obvious way, because the match() function is used in a number of different
5139 circumstances. The count starts from zero for each position in the subject
5140 string (for non-anchored patterns). The default limit is, for compatibility, a
5141 large number, namely 10 000 000. You can change this in two ways:
5143 (a) When configuring PCRE before making, you can use --with-match-limit=n
5144 to set a default value for the compiled library.
5146 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5147 a different value is set. See 45 below.
5149 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5151 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5152 of things that can be changed at compile time. The first argument specifies
5153 what is wanted and the second points to where the information is to be placed.
5154 The current list of available information is:
5158 The output is an integer that is set to one if UTF-8 support is available;
5159 otherwise it is set to zero.
5163 The output is an integer that it set to the value of the code that is used for
5164 newline. It is either LF (10) or CR (13).
5168 The output is an integer that contains the number of bytes used for internal
5169 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5173 The output is an integer that contains the threshold above which the POSIX
5174 interface uses malloc() for output vectors. See item 31 above.
5178 The output is an unsigned integer that contains the default limit of the number
5179 of match() calls in a pcre_exec() execution. See 42 above.
5181 44. pcretest has been upgraded by the addition of the -C option. This causes it
5182 to extract all the available output from the new pcre_config() function, and to
5183 output it. The program then exits immediately.
5185 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5186 order to support additional features. One way would have been to define
5187 pcre_exec2() (for example) with extra arguments, but this would not have been
5188 extensible, and would also have required all calls to the original function to
5189 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5190 is used for passing in "extra" data from pcre_study().
5192 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5193 contains the following fields:
5195 flags a bitmap indicating which of the following fields are set
5196 study_data opaque data from pcre_study()
5197 match_limit a way of specifying a limit on match() calls for a specific
5198 call to pcre_exec()
5199 callout_data data for callouts (see 49 below)
5201 The flag bits are also defined in pcre.h, and are
5207 The pcre_study() function now returns one of these new pcre_extra blocks, with
5208 the actual study data pointed to by the study_data field, and the
5209 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5210 before. That is, this change is entirely upwards-compatible and requires no
5211 change to existing code.
5213 If you want to pass in additional data to pcre_exec(), you can either place it
5214 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5215 block.
5217 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5218 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5219 times with different match limits, until it finds the minimum value needed for
5220 pcre_exec() to complete. The value is then output. This can be instructive; for
5221 most simple matches the number is quite small, but for pathological cases it
5222 gets very large very quickly.
5224 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5225 returns the size of the data block pointed to by the study_data field in a
5226 pcre_extra block, that is, the value that was passed as the argument to
5227 pcre_malloc() when PCRE was getting memory in which to place the information
5228 created by pcre_study(). The fourth argument should point to a size_t variable.
5229 pcretest has been extended so that this information is shown after a successful
5230 pcre_study() call when information about the compiled regex is being displayed.
5232 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5233 because what follows is always an absolute path. (Later: it turns out that this
5234 is more than cosmetic for MinGW, because it doesn't like empty path
5235 components.)
5237 49. Some changes have been made to the callout feature (see 28 above):
5239 (i) A callout function now has three choices for what it returns:
5241 0 => success, carry on matching
5242 > 0 => failure at this point, but backtrack if possible
5243 < 0 => serious error, return this value from pcre_exec()
5245 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5246 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5247 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5248 use by callout functions. It will never be used by PCRE itself.
5250 (ii) The pcre_extra structure (see 45 above) has a void * field called
5251 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5252 pcre_callout_block structure has a field of the same name. The contents of
5253 the field passed in the pcre_extra structure are passed to the callout
5254 function in the corresponding field in the callout block. This makes it
5255 easier to use the same callout-containing regex from multiple threads. For
5256 testing, the pcretest program has a new data escape
5258 \C*n pass the number n (may be negative) as callout_data
5260 If the callout function in pcretest receives a non-zero value as
5261 callout_data, it returns that value.
5263 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5264 there were some redundant $(CFLAGS) in commands that are now specified as
5265 $(LINK), which already includes $(CFLAGS).
5267 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5268 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5269 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5270 one-byte characters throughout. Note that case-insensitive matching applies
5271 only to characters whose values are less than 256. PCRE doesn't support the
5272 notion of cases for higher-valued characters.
5274 (i) A character class whose characters are all within 0-255 is handled as
5275 a bit map, and the map is inverted for negative classes. Previously, a
5276 character > 255 always failed to match such a class; however it should
5277 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5279 (ii) A negated character class with a single character < 255 is coded as
5280 "not this character" (OP_NOT). This wasn't working properly when the test
5281 character was multibyte, either singly or repeated.
5283 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5284 mode, for example: \x{100}{2,3}.
5286 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5287 singly or repeated) now correctly test multibyte characters. However,
5288 PCRE doesn't recognize any characters with values greater than 255 as
5289 digits, spaces, or word characters. Such characters always match \D, \S,
5290 and \W, and never match \d, \s, or \w.
5292 (v) Classes may now contain characters and character ranges with values
5293 greater than 255. For example: [ab\x{100}-\x{400}].
5295 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5296 PCRE in UTF-8 mode.
5298 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5299 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5300 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5301 value.)
5303 53. The single man page has become too large. I have therefore split it up into
5304 a number of separate man pages. These also give rise to individual HTML pages;
5305 these are now put in a separate directory, and there is an index.html page that
5306 lists them all. Some hyperlinking between the pages has been installed.
5308 54. Added convenience functions for handling named capturing parentheses.
5310 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5311 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5312 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5313 are faulted.
5315 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5316 calling configure. These values are used when compiling the dftables.c program
5317 which is run to generate the source of the default character tables. They
5318 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5319 you will need to set these values.
5321 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5324 Version 3.9 02-Jan-02
5325 ---------------------
5327 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5329 2. If --disable-static was given, the building process failed when trying to
5330 build pcretest and pcregrep. (For some reason it was using libtool to compile
5331 them, which is not right, as they aren't part of the library.)
5334 Version 3.8 18-Dec-01
5335 ---------------------
5337 1. The experimental UTF-8 code was completely screwed up. It was packing the
5338 bytes in the wrong order. How dumb can you get?
5341 Version 3.7 29-Oct-01
5342 ---------------------
5344 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5345 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5346 this didn't happen under Solaris 8, where I normally test things.
5348 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5349 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5350 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5351 this fix an hour or so after the initial 3.7 release.)
5354 Version 3.6 23-Oct-01
5355 ---------------------
5357 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5358 offsets passed as NULL with zero offset count.
5360 2. The config.guess and config.sub files had not been updated when I moved to
5361 the latest autoconf.
5364 Version 3.5 15-Aug-01
5365 ---------------------
5367 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5368 had been forgotten.
5370 2. By using declared but undefined structures, we can avoid using "void"
5371 definitions in pcre.h while keeping the internal definitions of the structures
5372 private.
5374 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5375 user point of view, this means that both static and shared libraries are built
5376 by default, but this can be individually controlled. More of the work of
5377 handling this static/shared cases is now inside libtool instead of PCRE's make
5378 file.
5380 4. The pcretest utility is now installed along with pcregrep because it is
5381 useful for users (to test regexs) and by doing this, it automatically gets
5382 relinked by libtool. The documentation has been turned into a man page, so
5383 there are now .1, .txt, and .html versions in /doc.
5385 5. Upgrades to pcregrep:
5386 (i) Added long-form option names like gnu grep.
5387 (ii) Added --help to list all options with an explanatory phrase.
5388 (iii) Added -r, --recursive to recurse into sub-directories.
5389 (iv) Added -f, --file to read patterns from a file.
5391 6. pcre_exec() was referring to its "code" argument before testing that
5392 argument for NULL (and giving an error if it was NULL).
5394 7. Upgraded Makefile.in to allow for compiling in a different directory from
5395 the source directory.
5397 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5398 options bits, the pointer it was passed was to an int instead of to an unsigned
5399 long int. This mattered only on 64-bit systems.
5401 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5402 generated) instead of pcre.in, which it its source. Also made the same change
5403 in several of the .c files.
5405 10. A new release of gcc defines printf() as a macro, which broke pcretest
5406 because it had an ifdef in the middle of a string argument for printf(). Fixed
5407 by using separate calls to printf().
5409 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5410 script, to force use of CR or LF instead of \n in the source. On non-Unix
5411 systems, the value can be set in config.h.
5413 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5414 absolute limit. Changed the text of the error message to make this clear, and
5415 likewise updated the man page.
5417 13. The limit of 99 on the number of capturing subpatterns has been removed.
5418 The new limit is 65535, which I hope will not be a "real" limit.
5421 Version 3.4 22-Aug-00
5422 ---------------------
5424 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5426 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5429 Version 3.3 01-Aug-00
5430 ---------------------
5432 1. If an octal character was given, but the value was greater than \377, it
5433 was not getting masked to the least significant bits, as documented. This could
5434 lead to crashes in some systems.
5436 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5437 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5439 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5440 These just pass their arguments on to (pcre_free)(), but they are provided
5441 because some uses of PCRE bind it to non-C systems that can call its functions,
5442 but cannot call free() or pcre_free() directly.
5444 4. Add "make test" as a synonym for "make check". Corrected some comments in
5445 the Makefile.
5447 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5448 Makefile.
5450 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5451 command called pgrep for grepping around the active processes.
5453 7. Added the beginnings of support for UTF-8 character strings.
5455 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5456 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5457 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5458 out for the ar command.)
5461 Version 3.2 12-May-00
5462 ---------------------
5464 This is purely a bug fixing release.
5466 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5467 of ZA. This was just one example of several cases that could provoke this bug,
5468 which was introduced by change 9 of version 2.00. The code for breaking
5469 infinite loops after an iteration that matches an empty string was't working
5470 correctly.
5472 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5473 when matched against abbab (for example). After matching an empty string, it
5474 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5475 caused it to match further down the string than it should.
5477 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5478 was there because it doesn't seem to be needed, and it causes trouble on some
5479 systems, as it is not a Standard C header. It has been removed.
5481 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5482 were reported on the Macintosh. The changes were from
5484 while ((c = *(++ptr)) != 0 && c != '\n');
5485 to
5486 while ((c = *(++ptr)) != 0 && c != '\n') ;
5488 Totally extraordinary, but if that's what it takes...
5490 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5491 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5492 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5493 assumes the way PCRE uses memmove() (always moving upwards).
5495 6. PCRE is being used in one environment where strchr() is not available. There
5496 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5497 faster code anyway.
5500 Version 3.1 09-Feb-00
5501 ---------------------
5503 The only change in this release is the fixing of some bugs in Makefile.in for
5504 the "install" target:
5506 (1) It was failing to install pcreposix.h.
5508 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5511 Version 3.0 01-Feb-00
5512 ---------------------
5514 1. Add support for the /+ modifier to perltest (to output $` like it does in
5515 pcretest).
5517 2. Add support for the /g modifier to perltest.
5519 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5520 matches null strings.
5522 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5523 pattern. Perl treats empty patterns specially - it reuses the most recent
5524 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5525 effect.
5527 5. The POSIX interface was broken in that it was just handing over the POSIX
5528 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5529 required a bigger vector, with some working space on the end. This means that
5530 the POSIX wrapper now has to get and free some memory, and copy the results.
5532 6. Added some simple autoconf support, placing the test data and the
5533 documentation in separate directories, re-organizing some of the
5534 information files, and making it build pcre-config (a GNU standard). Also added
5535 libtool support for building PCRE as a shared library, which is now the
5536 default.
5538 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5539 09 are not valid octal constants. Single digits will be used for minor values
5540 less than 10.
5542 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5543 existing programs that set these in the POSIX interface can use PCRE without
5544 modification.
5546 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5547 return all that pcre_info() returns, plus additional data. The pcre_info()
5548 function is retained for compatibility, but is considered to be obsolete.
5550 10. Added experimental recursion feature (?R) to handle one common case that
5551 Perl 5.6 will be able to do with (?p{...}).
5553 11. Added support for POSIX character classes like [:alpha:], which Perl is
5554 adopting.
5557 Version 2.08 31-Aug-99
5558 ----------------------
5560 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5561 trying to match at the startoffset position, but instead was moving forward to
5562 the next newline as if a previous match had failed.
5564 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5565 and could get into a loop if a null string was matched other than at the start
5566 of the subject.
5568 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5569 be distinguished at compile time, and for completeness also added PCRE_DATE.
5571 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5572 in GnuWin32 environments.
5575 Version 2.07 29-Jul-99
5576 ----------------------
5578 1. The documentation is now supplied in plain text form and HTML as well as in
5579 the form of man page sources.
5581 2. C++ compilers don't like assigning (void *) values to other pointer types.
5582 In particular this affects malloc(). Although there is no problem in Standard
5583 C, I've put in casts to keep C++ compilers happy.
5585 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5586 should be (const char *).
5588 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5589 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5590 However, I haven't made this a standard facility. The documentation doesn't
5591 mention it, and the Makefile doesn't support it.
5593 5. The Makefile now contains an "install" target, with editable destinations at
5594 the top of the file. The pcretest program is not installed.
5596 6. pgrep -V now gives the PCRE version number and date.
5598 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5599 causing the entire string to be ignored, instead of just the last character.
5601 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5602 non-matching string, it can take a very, very long time, even for strings of
5603 quite modest length, because of the nested recursion. PCRE now does better in
5604 some of these cases. It does this by remembering the last required literal
5605 character in the pattern, and pre-searching the subject to ensure it is present
5606 before running the real match. In other words, it applies a heuristic to detect
5607 some types of certain failure quickly, and in the above example, if presented
5608 with a string that has no trailing " it gives "no match" very quickly.
5610 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5611 other alternatives are tried instead.
5614 Version 2.06 09-Jun-99
5615 ----------------------
5617 1. Change pcretest's output for amount of store used to show just the code
5618 space, because the remainder (the data block) varies in size between 32-bit and
5619 64-bit systems.
5621 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5622 start matching at. This allows lookbehinds to work when searching for multiple
5623 occurrences in a string.
5625 3. Added additional options to pcretest for testing multiple occurrences:
5627 /+ outputs the rest of the string that follows a match
5628 /g loops for multiple occurrences, using the new startoffset argument
5629 /G loops for multiple occurrences by passing an incremented pointer
5631 4. PCRE wasn't doing the "first character" optimization for patterns starting
5632 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5633 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5634 the letter 'x'. On long subject strings, this gives a significant speed-up.
5637 Version 2.05 21-Apr-99
5638 ----------------------
5640 1. Changed the type of magic_number from int to long int so that it works
5641 properly on 16-bit systems.
5643 2. Fixed a bug which caused patterns starting with .* not to work correctly
5644 when the subject string contained newline characters. PCRE was assuming
5645 anchoring for such patterns in all cases, which is not correct because .* will
5646 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5647 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5648 must be retried after every newline in the subject.
5651 Version 2.04 18-Feb-99
5652 ----------------------
5654 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5655 computation of the store needed to hold the pattern was incorrect (too large).
5656 If such patterns were nested a few deep, this could multiply and become a real
5657 problem.
5659 2. Added /M option to pcretest to show the memory requirement of a specific
5660 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5662 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5663 compiled in such a way that the backtracking after subsequent failure was
5664 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5665 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5668 Version 2.03 02-Feb-99
5669 ----------------------
5671 1. Fixed typo and small mistake in man page.
5673 2. Added 4th condition (GPL supersedes if conflict) and created separate
5674 LICENCE file containing the conditions.
5676 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5677 Perl, that is the internal \ allows the delimiter to be included in the
5678 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5679 the final delimiter, add \ to the end of the pattern (to test the error).
5681 4. Added the convenience functions for extracting substrings after a successful
5682 match. Updated pcretest to make it able to test these functions.
5685 Version 2.02 14-Jan-99
5686 ----------------------
5688 1. Initialized the working variables associated with each extraction so that
5689 their saving and restoring doesn't refer to uninitialized store.
5691 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5692 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5693 fix the problem.
5695 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5696 calls, and wasn't printing the correct value for compiling calls. Increased the
5697 default value of LOOPREPEAT, and the number of significant figures in the
5698 times.
5700 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5702 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5703 a building problem on Windows NT with a FAT file system.
5706 Version 2.01 21-Oct-98
5707 ----------------------
5709 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5710 to character tables built by pcre_maketables() in the current locale. If NULL
5711 is passed, the default tables are used.
5714 Version 2.00 24-Sep-98
5715 ----------------------
5717 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5718 it any more.
5720 2. Allow quantification of (?>) groups, and make it work correctly.
5722 3. The first character computation wasn't working for (?>) groups.
5724 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5725 end of the subject) and add 5.005's \z, which really does match only at the
5726 very end of the subject.
5728 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5730 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5731 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5732 localized options. All options to pcre_study() were also removed.
5734 7. Add other new features from 5.005:
5736 $(?<= positive lookbehind
5737 $(?<! negative lookbehind
5738 (?imsx-imsx) added the unsetting capability
5739 such a setting is global if at outer level; local otherwise
5740 (?imsx-imsx:) non-capturing groups with option setting
5741 (?(cond)re|re) conditional pattern matching
5743 A backreference to itself in a repeated group matches the previous
5744 captured string.
5746 8. General tidying up of studying (both automatic and via "study")
5747 consequential on the addition of new assertions.
5749 9. As in 5.005, unlimited repeated groups that could match an empty substring
5750 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5751 runtime if any iteration does actually match an empty substring.
5753 10. Include the RunTest script in the distribution.
5755 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5756 discrepancies, some of which were old and were also with respect to 5.004. They
5757 have now been fixed.
5760 Version 1.09 28-Apr-98
5761 ----------------------
5763 1. A negated single character class followed by a quantifier with a minimum
5764 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5765 program crashes, or just wrong answers. This did not apply to negated classes
5766 containing more than one character, or to minima other than one.
5769 Version 1.08 27-Mar-98
5770 ----------------------
5772 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5774 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5775 latter must appear before anything that relies on it in the pattern.
5778 Version 1.07 16-Feb-98
5779 ----------------------
5781 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5782 repeat of a potentially empty string).
5785 Version 1.06 23-Jan-98
5786 ----------------------
5788 1. Added Markus Oberhumer's little patches for C++.
5790 2. Literal strings longer than 255 characters were broken.
5793 Version 1.05 23-Dec-97
5794 ----------------------
5796 1. Negated character classes containing more than one character were failing if
5797 PCRE_CASELESS was set at run time.
5800 Version 1.04 19-Dec-97
5801 ----------------------
5803 1. Corrected the man page, where some "const" qualifiers had been omitted.
5805 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5806 input syntax.
5808 3. Fixed memory leak which occurred when a regex with back references was
5809 matched with an offsets vector that wasn't big enough. The temporary memory
5810 that is used in this case wasn't being freed if the match failed.
5812 4. Tidied pcretest to ensure it frees memory that it gets.
5814 5. Temporary memory was being obtained in the case where the passed offsets
5815 vector was exactly big enough.
5817 6. Corrected definition of offsetof() from change 5 below.
5819 7. I had screwed up change 6 below and broken the rules for the use of
5820 setjmp(). Now fixed.
5823 Version 1.03 18-Dec-97
5824 ----------------------
5826 1. A erroneous regex with a missing opening parenthesis was correctly
5827 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5828 on some systems.
5830 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5831 it was reported that one broken compiler failed on the former because "code" is
5832 also an independent variable.
5834 3. The erroneous regex a[]b caused an array overrun reference.
5836 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5837 fail on data ending with that character. (It was going on too far, and checking
5838 the next character, typically a binary zero.) This was specific to the
5839 optimized code for single-character negative classes.
5841 5. Added a contributed patch from the TIN world which does the following:
5843 + Add an undef for memmove, in case the the system defines a macro for it.
5845 + Add a definition of offsetof(), in case there isn't one. (I don't know
5846 the reason behind this - offsetof() is part of the ANSI standard - but
5847 it does no harm).
5849 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5850 most of the places where whitespace preceded '#'. I have given up and
5851 allowed the remaining 2 cases to be at the margin.
5853 + Rename some variables in pcre to eliminate shadowing. This seems very
5854 pedantic, but does no harm, of course.
5856 6. Moved the call to setjmp() into its own function, to get rid of warnings
5857 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5859 7. Constructs such as \d{8,} were compiling into the equivalent of
5860 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5861 outcome, but in this particular case used more store than had been allocated,
5862 which caused the bug to be discovered because it threw up an internal error.
5864 8. The debugging code in both pcre and pcretest for outputting the compiled
5865 form of a regex was going wrong in the case of back references followed by
5866 curly-bracketed repeats.
5869 Version 1.02 12-Dec-97
5870 ----------------------
5872 1. Typos in pcre.3 and comments in the source fixed.
5874 2. Applied a contributed patch to get rid of places where it used to remove
5875 'const' from variables, and fixed some signed/unsigned and uninitialized
5876 variable warnings.
5878 3. Added the "runtest" target to Makefile.
5880 4. Set default compiler flag to -O2 rather than just -O.
5883 Version 1.01 19-Nov-97
5884 ----------------------
5886 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5887 like /([ab]*)*/, that is, for classes with more than one character in them.
5889 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5890 as /((?>a*))*/ (a PCRE_EXTRA facility).
5893 Version 1.00 18-Nov-97
5894 ----------------------
5896 1. Added compile-time macros to support systems such as SunOS4 which don't have
5897 memmove() or strerror() but have other things that can be used instead.
5899 2. Arranged that "make clean" removes the executables.
5902 Version 0.99 27-Oct-97
5903 ----------------------
5905 1. Fixed bug in code for optimizing classes with only one character. It was
5906 initializing a 32-byte map regardless, which could cause it to run off the end
5907 of the memory it had got.
5909 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5912 Version 0.98 22-Oct-97
5913 ----------------------
5915 1. Fixed bug in code for handling temporary memory usage when there are more
5916 back references than supplied space in the ovector. This could cause segfaults.
5919 Version 0.97 21-Oct-97
5920 ----------------------
5922 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5924 2. Optimized negated single characters not to use a bit map.
5926 3. Brought error texts together as macro definitions; clarified some of them;
5927 fixed one that was wrong - it said "range out of order" when it meant "invalid
5928 escape sequence".
5930 4. Changed some char * arguments to const char *.
5932 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5934 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5935 pcretest.
5938 Version 0.96 16-Oct-97
5939 ----------------------
5941 1. Added a simple "pgrep" utility to the distribution.
5943 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5944 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5945 where "ddd" means "one or more decimal digits".
5947 3. Fixed serious bug. If a pattern had a back reference, but the call to
5948 pcre_exec() didn't supply a large enough ovector to record the related
5949 identifying subpattern, the match always failed. PCRE now remembers the number
5950 of the largest back reference, and gets some temporary memory in which to save
5951 the offsets during matching if necessary, in order to ensure that
5952 backreferences always work.
5954 4. Increased the compatibility with Perl in a number of ways:
5956 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5957 to request this handling. The option can be set at compile or exec time.
5959 (b) $ matches before a terminating newline by default; an option
5960 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5961 mode). The option can be set at compile or exec time.
5963 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5964 the same as Perl's. If the decimal number it represents is less than 10
5965 or there aren't that many previous left capturing parentheses, an octal
5966 escape is read. Inside a character class, it's always an octal escape,
5967 even if it is a single digit.
5969 (d) An escaped but undefined alphabetic character is taken as a literal,
5970 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5971 escapes.
5973 (e) {0} is now permitted. (The previous item is removed from the compiled
5974 pattern).
5976 5. Changed all the names of code files so that the basic parts are no longer
5977 than 10 characters, and abolished the teeny "globals.c" file.
5979 6. Changed the handling of character classes; they are now done with a 32-byte
5980 bit map always.
5982 7. Added the -d and /D options to pcretest to make it possible to look at the
5983 internals of compilation without having to recompile pcre.
5986 Version 0.95 23-Sep-97
5987 ----------------------
5989 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5990 \x20 at the start of a run of normal characters. These were being treated as
5991 real characters, instead of the source characters being re-checked.
5994 Version 0.94 18-Sep-97
5995 ----------------------
5997 1. The functions are now thread-safe, with the caveat that the global variables
5998 containing pointers to malloc() and free() or alternative functions are the
5999 same for all threads.
6001 2. Get pcre_study() to generate a bitmap of initial characters for non-
6002 anchored patterns when this is possible, and use it if passed to pcre_exec().
6005 Version 0.93 15-Sep-97
6006 ----------------------
6008 1. /(b)|(:+)/ was computing an incorrect first character.
6010 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
6011 but not actually doing anything yet.
6013 3. Treat "-" characters in classes that cannot be part of ranges as literals,
6014 as Perl does (e.g. [-az] or [az-]).
6016 4. Set the anchored flag if a branch starts with .* or .*? because that tests
6017 all possible positions.
6019 5. Split up into different modules to avoid including unneeded functions in a
6020 compiled binary. However, compile and exec are still in one module. The "study"
6021 function is split off.
6023 6. The character tables are now in a separate module whose source is generated
6024 by an auxiliary program - but can then be edited by hand if required. There are
6025 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
6026 toupper() in the code.
6028 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
6029 make them global. Abolish the function for setting them, as the caller can now
6030 set them directly.
6033 Version 0.92 11-Sep-97
6034 ----------------------
6036 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
6037 (e.g. /a{1,3}/) was broken (I mis-optimized it).
6039 2. Caseless matching was not working in character classes if the characters in
6040 the pattern were in upper case.
6042 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
6044 4. Make PCRE_ANCHORED public and accept as a compile option.
6046 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
6047 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
6048 pass them.
6050 6. Give an error if bad option bits passed at compile or run time.
6052 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
6053 pcretest to cause it to pass that flag.
6055 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
6056 options, and the first character, if set.
6058 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
6061 Version 0.91 10-