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 8.39 14-June-2016
8 -------------------------
10 1. If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
11 an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
12 misbehaved. This bug was found by the LLVM fuzzer.
14 2. Similar to the above, if an isolated \E was present between an item and its
15 qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
16 bug was found by the LLVM fuzzer.
18 3. Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
19 working correctly in UCP mode.
21 4. The POSIX wrapper function regexec() crashed if the option REG_STARTEND
22 was set when the pmatch argument was NULL. It now returns REG_INVARG.
24 5. Allow for up to 32-bit numbers in the ordin() function in pcregrep.
26 6. An empty \Q\E sequence between an item and its qualifier caused
27 pcre_compile() to misbehave when auto callouts were enabled. This bug was
28 found by the LLVM fuzzer.
30 7. If a pattern that was compiled with PCRE_EXTENDED started with white
31 space or a #-type comment that was followed by (?-x), which turns off
32 PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
33 pcre_compile() assumed that (?-x) applied to the whole pattern and
34 consequently mis-compiled it. This bug was found by the LLVM fuzzer.
36 8. A call of pcre_copy_named_substring() for a named substring whose number
37 was greater than the space in the ovector could cause a crash.
39 9. Yet another buffer overflow bug involved duplicate named groups with a
40 group that reset capture numbers (compare 8.38/7 below). Once again, I have
41 just allowed for more memory, even if not needed. (A proper fix is
42 implemented in PCRE2, but it involves a lot of refactoring.)
44 10. pcre_get_substring_list() crashed if the use of \K in a match caused the
45 start of the match to be earlier than the end.
47 11. Migrating appropriate PCRE2 JIT improvements to PCRE.
49 12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
50 assertion, caused pcretest to generate incorrect output, and also to read
51 uninitialized memory (detected by ASAN or valgrind).
53 13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
54 nested set of parentheses of sufficient size caused an overflow of the
55 compiling workspace (which was diagnosed, but of course is not desirable).
57 14. And yet another buffer overflow bug involving duplicate named groups, this
58 time nested, with a nested back reference. Yet again, I have just allowed
59 for more memory, because anything more needs all the refactoring that has
60 been done for PCRE2. An example pattern that provoked this bug is:
61 /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
62 registered as CVE-2016-1283.
64 15. pcretest went into a loop if global matching was requested with an ovector
65 size less than 2. It now gives an error message. This bug was found by
68 16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
69 ("assertion expected") when (?(?C) was not followed by an opening
72 17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
73 actually affect anything, by sheer luck.
75 18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
78 19. Modified the RunTest script to incorporate a valgrind suppressions file so
79 that certain errors, provoked by the SSE2 instruction set when JIT is used,
82 20. A racing condition is fixed in JIT reported by Mozilla.
84 21. Minor code refactor to avoid "array subscript is below array bounds"
87 22. Minor code refactor to avoid "left shift of negative number" warning.
89 23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
92 24. Refactor to avoid compiler warnings in pcrecpp.cc.
94 25. Refactor to fix a typo in pcre_jit_test.c
96 26. Patch to support compiling pcrecpp.cc with Intel compiler.
99 Version 8.38 23-November-2015
100 -----------------------------
102 1. If a group that contained a recursive back reference also contained a
103 forward reference subroutine call followed by a non-forward-reference
104 subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
105 compile correct code, leading to undefined behaviour or an internally
106 detected error. This bug was discovered by the LLVM fuzzer.
108 2. Quantification of certain items (e.g. atomic back references) could cause
109 incorrect code to be compiled when recursive forward references were
110 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
111 This bug was discovered by the LLVM fuzzer.
113 3. A repeated conditional group whose condition was a reference by name caused
114 a buffer overflow if there was more than one group with the given name.
115 This bug was discovered by the LLVM fuzzer.
117 4. A recursive back reference by name within a group that had the same name as
118 another group caused a buffer overflow. For example:
119 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
121 5. A forward reference by name to a group whose number is the same as the
122 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
123 a buffer overflow at compile time. This bug was discovered by the LLVM
126 6. A lookbehind assertion within a set of mutually recursive subpatterns could
127 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
129 7. Another buffer overflow bug involved duplicate named groups with a
130 reference between their definition, with a group that reset capture
131 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
132 fixed by always allowing for more memory, even if not needed. (A proper fix
133 is implemented in PCRE2, but it involves more refactoring.)
135 8. There was no check for integer overflow in subroutine calls such as (?123).
137 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
138 being treated as a literal 'l' instead of causing an error.
140 10. There was a buffer overflow if pcre_exec() was called with an ovector of
141 size 1. This bug was found by american fuzzy lop.
143 11. If a non-capturing group containing a conditional group that could match
144 an empty string was repeated, it was not identified as matching an empty
145 string itself. For example: /^(?:(?(1)x|)+)+$()/.
147 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
148 \a and \e in test subject lines.
150 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
151 instead of the EBCDIC value.
153 14. The handling of \c in an EBCDIC environment has been revised so that it is
154 now compatible with the specification in Perl's perlebcdic page.
156 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
157 ASCII/Unicode. This has now been added to the list of characters that are
158 recognized as white space in EBCDIC.
160 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
161 error (correctly) when used outside a class, but did not give an error
164 17. \h within a class was incorrectly compiled in EBCDIC environments.
166 18. A pattern with an unmatched closing parenthesis that contained a backward
167 assertion which itself contained a forward reference caused buffer
168 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
170 19. JIT should return with error when the compiled pattern requires more stack
171 space than the maximum.
173 20. A possessively repeated conditional group that could match an empty string,
174 for example, /(?(R))*+/, was incorrectly compiled.
176 21. Fix infinite recursion in the JIT compiler when certain patterns such as
177 /(?:|a|){100}x/ are analysed.
179 22. Some patterns with character classes involving [: and \\ were incorrectly
180 compiled and could cause reading from uninitialized memory or an incorrect
183 23. Pathological patterns containing many nested occurrences of [: caused
184 pcre_compile() to run for a very long time.
186 24. A conditional group with only one branch has an implicit empty alternative
187 branch and must therefore be treated as potentially matching an empty
190 25. If (?R was followed by - or + incorrect behaviour happened instead of a
193 26. Arrange to give up on finding the minimum matching length for overly
196 27. Similar to (4) above: in a pattern with duplicated named groups and an
197 occurrence of (?| it is possible for an apparently non-recursive back
198 reference to become recursive if a later named group with the relevant
199 number is encountered. This could lead to a buffer overflow. Wen Guanxing
200 from Venustech ADLAB discovered this bug.
202 28. If pcregrep was given the -q option with -c or -l, or when handling a
203 binary file, it incorrectly wrote output to stdout.
205 29. The JIT compiler did not restore the control verb head in case of *THEN
206 control verbs. This issue was found by Karl Skomski with a custom LLVM
209 30. Error messages for syntax errors following \g and \k were giving inaccurate
210 offsets in the pattern.
212 31. Added a check for integer overflow in conditions (?(<digits>) and
213 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
216 32. Handling recursive references such as (?2) when the reference is to a group
217 later in the pattern uses code that is very hacked about and error-prone.
218 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
219 give an internal error if it is obvious that compiling has gone wrong.
221 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
222 This issue was found by Karl Skomski with a custom LLVM fuzzer.
224 34. The JIT compiler should restore the control chain for empty possessive
225 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
227 35. Match limit check added to JIT recursion. This issue was found by Karl
228 Skomski with a custom LLVM fuzzer.
230 36. Yet another case similar to 27 above has been circumvented by an
231 unconditional allocation of extra memory. This issue is fixed "properly" in
232 PCRE2 by refactoring the way references are handled. Wen Guanxing
233 from Venustech ADLAB discovered this bug.
235 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
236 with a custom LLVM fuzzer.
238 38. Fixed a corner case of range optimization in JIT.
240 39. An incorrect error "overran compiling workspace" was given if there were
241 exactly enough group forward references such that the last one extended
242 into the workspace safety margin. The next one would have expanded the
243 workspace. The test for overflow was not including the safety margin.
245 40. A match limit issue is fixed in JIT which was found by Karl Skomski
246 with a custom LLVM fuzzer.
248 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
249 work under Windows. (Why has it taken so long for anyone to notice?)
251 42. In a character class such as [\W\p{Any}] where both a negative-type escape
252 ("not a word character") and a property escape were present, the property
253 escape was being ignored.
255 43. Fix crash caused by very long (*MARK) or (*THEN) names.
257 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
258 by a single ASCII character in a class item, was incorrectly compiled in
259 UCP mode. The POSIX class got lost, but only if the single character
262 45. [:punct:] in UCP mode was matching some characters in the range 128-255
263 that should not have been matched.
265 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
266 class, all characters with code points greater than 255 are in the class.
267 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
268 such as \w are turned into Unicode properties), wide characters were not
269 correctly handled, and could fail to match.
272 Version 8.37 28-April-2015
273 --------------------------
275 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
276 for those parentheses to be closed with whatever has been captured so far.
277 However, it was failing to mark any other groups between the hightest
278 capture so far and the currrent group as "unset". Thus, the ovector for
279 those groups contained whatever was previously there. An example is the
280 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
282 2. If an assertion condition was quantified with a minimum of zero (an odd
283 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
285 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
286 unrecognized modifier, a crash could occur.
288 4. An attempt to do global matching in pcretest with a zero-length ovector
291 5. Fixed a memory leak during matching that could occur for a subpattern
292 subroutine call (recursive or otherwise) if the number of captured groups
293 that had to be saved was greater than ten.
295 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
296 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
297 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
299 7. A UTF pattern containing a "not" match of a non-ASCII character and a
300 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
302 8. When a pattern is compiled, it remembers the highest back reference so that
303 when matching, if the ovector is too small, extra memory can be obtained to
304 use instead. A conditional subpattern whose condition is a check on a
305 capture having happened, such as, for example in the pattern
306 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
307 setting the highest backreference number. This mattered only if pcre_exec()
308 was called with an ovector that was too small to hold the capture, and there
309 was no other kind of back reference (a situation which is probably quite
310 rare). The effect of the bug was that the condition was always treated as
311 FALSE when the capture could not be consulted, leading to a incorrect
312 behaviour by pcre_exec(). This bug has been fixed.
314 9. A reference to a duplicated named group (either a back reference or a test
315 for being set in a conditional) that occurred in a part of the pattern where
316 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
317 to be incorrectly calculated, leading to overwriting.
319 10. A mutually recursive set of back references such as (\2)(\1) caused a
320 segfault at study time (while trying to find the minimum matching length).
321 The infinite loop is now broken (with the minimum length unset, that is,
324 11. If an assertion that was used as a condition was quantified with a minimum
325 of zero, matching went wrong. In particular, if the whole group had
326 unlimited repetition and could match an empty string, a segfault was
327 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
328 assertions to be quantified, but not if they are being used as conditions,
329 so the above pattern is faulted by Perl. PCRE has now been changed so that
330 it also rejects such patterns.
332 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
333 failed to allow the zero-repeat case if pcre2_exec() was called with an
334 ovector too small to capture the group.
336 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
337 Red Hat Product Security:
339 (a) A crash if /K and /F were both set with the option to save the compiled
342 (b) Another crash if the option to print captured substrings in a callout
343 was combined with setting a null ovector, for example \O\C+ as a subject
346 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
347 forward reference repeated a large (but limited) number of times within a
348 repeated outer group that has a zero minimum quantifier, caused incorrect
349 code to be compiled, leading to the error "internal error:
350 previously-checked referenced subpattern not found" when an incorrect
351 memory address was read. This bug was reported as "heap overflow",
352 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
355 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
356 call within a group that also contained a recursive back reference caused
357 incorrect code to be compiled. This bug was reported as "heap overflow",
358 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
359 number CVE-2015-2326.
361 24. Computing the size of the JIT read-only data in advance has been a source
362 of various issues, and new ones are still appear unfortunately. To fix
363 existing and future issues, size computation is eliminated from the code,
364 and replaced by on-demand memory allocation.
366 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
367 adjacent to the end of the range, and the range contained characters with
368 more than one other case, caused incorrect behaviour when compiled in UTF
369 mode. In that example, the range a-j was left out of the class.
371 26. Fix JIT compilation of conditional blocks, which assertion
372 is converted to (*FAIL). E.g: /(?(?!))/.
374 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
375 discovered by the LLVM fuzzer.
377 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
378 when this assertion was used as a condition, for example (?(?!)a|b). In
379 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
380 error about an unsupported item.
382 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
383 possessification code could take exponential time to complete. A recursion
384 depth limit of 1000 has been imposed to limit the resources used by this
387 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
388 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
389 because \S ensures they are all in the class. The code for doing this was
390 interacting badly with the code for computing the amount of space needed to
391 compile the pattern, leading to a buffer overflow. This bug was discovered
394 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
395 other kinds of group caused stack overflow at compile time. This bug was
396 discovered by the LLVM fuzzer.
398 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
399 between a subroutine call and its quantifier was incorrectly compiled,
400 leading to buffer overflow or other errors. This bug was discovered by the
403 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
404 assertion after (?(. The code was failing to check the character after
405 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
406 was discovered by the LLVM fuzzer.
408 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
409 a fixed maximum following a group that contains a subroutine reference was
410 incorrectly compiled and could trigger buffer overflow. This bug was
411 discovered by the LLVM fuzzer.
413 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
414 caused a stack overflow instead of the diagnosis of a non-fixed length
415 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
417 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
418 (e.g. /(?<=\Ka)/) could make pcregrep loop.
420 37. There was a similar problem to 36 in pcretest for global matches.
422 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
423 and a subsequent item in the pattern caused a non-match, backtracking over
424 the repeated \X did not stop, but carried on past the start of the subject,
425 causing reference to random memory and/or a segfault. There were also some
426 other cases where backtracking after \C could crash. This set of bugs was
427 discovered by the LLVM fuzzer.
429 39. The function for finding the minimum length of a matching string could take
430 a very long time if mutual recursion was present many times in a pattern,
431 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
432 method has been implemented. This infelicity was discovered by the LLVM
435 40. Static linking against the PCRE library using the pkg-config module was
436 failing on missing pthread symbols.
439 Version 8.36 26-September-2014
440 ------------------------------
442 1. Got rid of some compiler warnings in the C++ modules that were shown up by
443 -Wmissing-field-initializers and -Wunused-parameter.
445 2. The tests for quantifiers being too big (greater than 65535) were being
446 applied after reading the number, and stupidly assuming that integer
447 overflow would give a negative number. The tests are now applied as the
450 3. Tidy code in pcre_exec.c where two branches that used to be different are
453 4. The JIT compiler did not generate match limit checks for certain
454 bracketed expressions with quantifiers. This may lead to exponential
455 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
456 issue should be resolved now.
458 5. Fixed an issue, which occures when nested alternatives are optimized
461 6. Inserted two casts and changed some ints to size_t in the light of some
462 reported 64-bit compiler warnings (Bugzilla 1477).
464 7. Fixed a bug concerned with zero-minimum possessive groups that could match
465 an empty string, which sometimes were behaving incorrectly in the
466 interpreter (though correctly in the JIT matcher). This pcretest input is
469 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
470 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
472 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
473 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
474 for an empty string was breaking the inner loop and carrying on at a lower
475 level, when possessive repeated groups should always return to a higher
476 level as they have no backtrack points in them. The empty string test now
477 occurs at the outer level.
479 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
480 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
482 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
485 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
486 to hold a pattern to be incorrectly computed (too small) when there were
487 named back references to duplicated names. This could cause "internal
488 error: code overflow" or "double free or corruption" or other memory
491 11. When named subpatterns had the same prefixes, back references could be
492 confused. For example, in this pattern:
494 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
496 the reference to 'Name' was incorrectly treated as a reference to a
499 12. A pattern such as /^s?c/mi8 where the optional character has more than
500 one "other case" was incorrectly compiled such that it would only try to
501 match starting at "c".
503 13. When a pattern starting with \s was studied, VT was not included in the
504 list of possible starting characters; this should have been part of the
507 14. If a character class started [\Qx]... where x is any character, the class
508 was incorrectly terminated at the ].
510 15. If a pattern that started with a caseless match for a character with more
511 than one "other case" was studied, PCRE did not set up the starting code
512 unit bit map for the list of possible characters. Now it does. This is an
513 optimization improvement, not a bug fix.
515 16. The Unicode data tables have been updated to Unicode 7.0.0.
517 17. Fixed a number of memory leaks in pcregrep.
519 18. Avoid a compiler warning (from some compilers) for a function call with
520 a cast that removes "const" from an lvalue by using an intermediate
521 variable (to which the compiler does not object).
523 19. Incorrect code was compiled if a group that contained an internal recursive
524 back reference was optional (had quantifier with a minimum of zero). This
525 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
526 caused segmentation faults because of stack overflows at compile time.
528 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
529 group that is quantified with an indefinite repeat, caused a compile-time
530 loop which used up all the system stack and provoked a segmentation fault.
531 This was not the same bug as 19 above.
533 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
534 Patch by Mike Frysinger.
537 Version 8.35 04-April-2014
538 --------------------------
540 1. A new flag is set, when property checks are present in an XCLASS.
541 When this flag is not set, PCRE can perform certain optimizations
542 such as studying these XCLASS-es.
544 2. The auto-possessification of character sets were improved: a normal
545 and an extended character set can be compared now. Furthermore
546 the JIT compiler optimizes more character set checks.
548 3. Got rid of some compiler warnings for potentially uninitialized variables
549 that show up only when compiled with -O2.
551 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
552 of a match later then the end of the match. The pcretest program was not
553 handling the case sensibly - it was outputting from the start to the next
554 binary zero. It now reports this situation in a message, and outputs the
555 text from the end to the start.
557 5. Fast forward search is improved in JIT. Instead of the first three
558 characters, any three characters with fixed position can be searched.
559 Search order: first, last, middle.
561 6. Improve character range checks in JIT. Characters are read by an inprecise
562 function now, which returns with an unknown value if the character code is
563 above a certain threshold (e.g: 256). The only limitation is that the value
564 must be bigger than the threshold as well. This function is useful when
565 the characters above the threshold are handled in the same way.
567 7. The macros whose names start with RAWUCHAR are placeholders for a future
568 mode in which only the bottom 21 bits of 32-bit data items are used. To
569 make this more memorable for those maintaining the code, the names have
570 been changed to start with UCHAR21, and an extensive comment has been added
573 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
574 to the export list in Makefile.am (they were accidentally omitted from the
577 9. The informational output from pcretest used the phrase "starting byte set"
578 which is inappropriate for the 16-bit and 32-bit libraries. As the output
579 for "first char" and "need char" really means "non-UTF-char", I've changed
580 "byte" to "char", and slightly reworded the output. The documentation about
581 these values has also been (I hope) clarified.
583 10. Another JIT related optimization: use table jumps for selecting the correct
584 backtracking path, when more than four alternatives are present inside a
587 11. Empty match is not possible, when the minimum length is greater than zero,
588 and there is no \K in the pattern. JIT should avoid empty match checks in
591 12. In a caseless character class with UCP support, when a character with more
592 than one alternative case was not the first character of a range, not all
593 the alternative cases were added to the class. For example, s and \x{17f}
594 are both alternative cases for S: the class [RST] was handled correctly,
597 13. The configure.ac file always checked for pthread support when JIT was
598 enabled. This is not used in Windows, so I have put this test inside a
599 check for the presence of windows.h (which was already tested for).
601 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
602 The algorithm provides a way to skip certain starting offsets, and usually
603 faster than linear prefix searches.
605 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
606 as for 'fr_FR' and 'french'. For some reason, however, it then used the
607 Windows-specific input and output files, which have 'french' screwed in.
608 So this could never have worked. One of the problems with locales is that
609 they aren't always the same. I have now updated RunTest so that it checks
610 the output of the locale test (test 3) against three different output
611 files, and it allows the test to pass if any one of them matches. With luck
612 this should make the test pass on some versions of Solaris where it was
613 failing. Because of the uncertainty, the script did not used to stop if
614 test 3 failed; it now does. If further versions of a French locale ever
615 come to light, they can now easily be added.
617 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
618 there was a message during ./configure, but it did not stop. This now
619 provokes an error. The invalid example in README has been corrected.
620 If a value less than the minimum is given, the minimum value has always
621 been used, but now a warning is given.
623 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
624 was a bug in the test system, which is now fixed. Also, the list of various
625 configurations that are tested for each release did not have one with both
626 16/32 bits and --enable-bar-anycrlf. It now does.
628 18. pcretest was missing "-C bsr" for displaying the \R default setting.
630 19. Little endian PowerPC systems are supported now by the JIT compiler.
632 20. The fast forward newline mechanism could enter to an infinite loop on
633 certain invalid UTF-8 input. Although we don't support these cases
634 this issue can be fixed by a performance optimization.
636 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
637 not take account if existing stack usage. There is now a new global
638 variable called pcre_stack_guard that can be set to point to an external
639 function to check stack availability. It is called at the start of
640 processing every parenthesized group.
642 22. A typo in the code meant that in ungreedy mode the max/min qualifier
643 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
646 23. When UTF was disabled, the JIT program reported some incorrect compile
647 errors. These messages are silenced now.
649 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
652 25. Change all the temporary files used in RunGrepTest to be different to those
653 used by RunTest so that the tests can be run simultaneously, for example by
657 Version 8.34 15-December-2013
658 -----------------------------
660 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
661 executable memory. Patch inspired by Carsten Klein.
663 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
664 this macro is never tested and has no effect, because the work to support
665 coverage involves only compiling and linking options and special targets in
666 the Makefile. The comment in config.h implied that defining the macro would
667 enable coverage support, which is totally false. There was also support for
668 setting this macro in the CMake files (my fault, I just copied it from
669 configure). SUPPORT_GCOV has now been removed.
671 3. Make a small performance improvement in strlen16() and strlen32() in
674 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
675 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
676 these cases). There was also one in pcretest.c.
678 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
680 6. In UTF mode, the code for checking whether a group could match an empty
681 string (which is used for indefinitely repeated groups to allow for
682 breaking an infinite loop) was broken when the group contained a repeated
683 negated single-character class with a character that occupied more than one
684 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
685 UTF-8 mode). The effect was undefined: the group might or might not be
686 deemed as matching an empty string, or the program might have crashed.
688 7. The code for checking whether a group could match an empty string was not
689 recognizing that \h, \H, \v, \V, and \R must match a character.
691 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
692 an empty string. If it can, pcretest shows this in its information output.
694 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
695 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
696 matched by pcre_exec() without using JIT:
698 (a) If the rest of the pattern did not match after a maximal run of
699 grapheme clusters, the code for backing up to try with fewer of them
700 did not always back up over a full grapheme when characters that do not
701 have the modifier quality were involved, e.g. Hangul syllables.
703 (b) If the match point in a subject started with modifier character, and
704 there was no match, the code could incorrectly back up beyond the match
705 point, and potentially beyond the first character in the subject,
706 leading to a segfault or an incorrect match result.
708 10. A conditional group with an assertion condition could lead to PCRE
709 recording an incorrect first data item for a match if no other first data
710 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
711 first data item, and therefore matched "ca" after "c" instead of at the
714 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
715 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
717 12. The source of pcregrep now includes z/OS-specific code so that it can be
718 compiled for z/OS as part of the special z/OS distribution.
720 13. Added the -T and -TM options to pcretest.
722 14. The code in pcre_compile.c for creating the table of named capturing groups
723 has been refactored. Instead of creating the table dynamically during the
724 actual compiling pass, the information is remembered during the pre-compile
725 pass (on the stack unless there are more than 20 named groups, in which
726 case malloc() is used) and the whole table is created before the actual
727 compile happens. This has simplified the code (it is now nearly 150 lines
728 shorter) and prepared the way for better handling of references to groups
729 with duplicate names.
731 15. A back reference to a named subpattern when there is more than one of the
732 same name now checks them in the order in which they appear in the pattern.
733 The first one that is set is used for the reference. Previously only the
734 first one was inspected. This change makes PCRE more compatible with Perl.
736 16. Unicode character properties were updated from Unicode 6.3.0.
738 17. The compile-time code for auto-possessification has been refactored, based
739 on a patch by Zoltan Herczeg. It now happens after instead of during
740 compilation. The code is cleaner, and more cases are handled. The option
741 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
742 options in pcretest are provided to set it. It can also be set by
743 (*NO_AUTO_POSSESS) at the start of a pattern.
745 18. The character VT has been added to the default ("C" locale) set of
746 characters that match \s and are generally treated as white space,
747 following this same change in Perl 5.18. There is now no difference between
748 "Perl space" and "POSIX space". Whether VT is treated as white space in
749 other locales depends on the locale.
751 19. The code for checking named groups as conditions, either for being set or
752 for being recursed, has been refactored (this is related to 14 and 15
753 above). Processing unduplicated named groups should now be as fast at
754 numerical groups, and processing duplicated groups should be faster than
757 20. Two patches to the CMake build system, by Alexander Barkov:
759 (1) Replace the "source" command by "." in CMakeLists.txt because
760 "source" is a bash-ism.
762 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
763 without these the CMake build does not work on Solaris.
765 21. Perl has changed its handling of \8 and \9. If there is no previously
766 encountered capturing group of those numbers, they are treated as the
767 literal characters 8 and 9 instead of a binary zero followed by the
768 literals. PCRE now does the same.
770 22. Following Perl, added \o{} to specify codepoints in octal, making it
771 possible to specify values greater than 0777 and also making them
774 23. Perl now gives an error for missing closing braces after \x{... instead of
775 treating the string as literal. PCRE now does the same.
777 24. RunTest used to grumble if an inappropriate test was selected explicitly,
778 but just skip it when running all tests. This make it awkward to run ranges
779 of tests when one of them was inappropriate. Now it just skips any
780 inappropriate tests, as it always did when running all tests.
782 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
783 character types such as \d or \w, too many callouts were inserted, and the
784 data that they returned was rubbish.
786 26. In UCP mode, \s was not matching two of the characters that Perl matches,
787 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
788 were matched by \h. The code has now been refactored so that the lists of
789 the horizontal and vertical whitespace characters used for \h and \v (which
790 are defined only in one place) are now also used for \s.
792 27. Add JIT support for the 64 bit TileGX architecture.
793 Patch by Jiong Wang (Tilera Corporation).
795 28. Possessive quantifiers for classes (both explicit and automatically
796 generated) now use special opcodes instead of wrapping in ONCE brackets.
798 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
799 (because it's meaningless), this was not happening when PCRE_CASELESS was
800 set. Not wrong, but inefficient.
802 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
803 properties for \w, \d, etc) is present in a test regex. Otherwise if the
804 test contains no characters greater than 255, Perl doesn't realise it
805 should be using Unicode semantics.
807 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
808 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
809 does in Unicode mode.
811 32. Added the "forbid" facility to pcretest so that putting tests into the
812 wrong test files can sometimes be quickly detected.
814 33. There is now a limit (default 250) on the depth of nesting of parentheses.
815 This limit is imposed to control the amount of system stack used at compile
816 time. It can be changed at build time by --with-parens-nest-limit=xxx or
817 the equivalent in CMake.
819 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
820 errors. Perl warns for these when in warning mode, but PCRE has no facility
823 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
824 However, this was not working for (?!) because it is optimized to (*FAIL),
825 for which PCRE does not allow quantifiers. The optimization is now disabled
826 when a quantifier follows (?!). I can't see any use for this, but it makes
829 36. Perl no longer allows group names to start with digits, so I have made this
830 change also in PCRE. It simplifies the code a bit.
832 37. In extended mode, Perl ignores spaces before a + that indicates a
833 possessive quantifier. PCRE allowed a space before the quantifier, but not
834 before the possessive +. It now does.
836 38. The use of \K (reset reported match start) within a repeated possessive
837 group such as (a\Kb)*+ was not working.
839 40. Document that the same character tables must be used at compile time and
840 run time, and that the facility to pass tables to pcre_exec() and
841 pcre_dfa_exec() is for use only with saved/restored patterns.
843 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
844 features for Builds with MSVC:
846 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
847 builds). The .rc files can be used to set FileDescription and many other
850 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
851 This allows higher-level build scripts which want .pdb files to avoid
852 hard-coding the exact files needed."
854 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
855 mean "start of word" and "end of word", respectively, as a transition aid.
857 43. A minimizing repeat of a class containing codepoints greater than 255 in
858 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
859 compiled to use the heap for recursion.
861 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
865 Version 8.33 28-May-2013
866 ------------------------
868 1. Added 'U' to some constants that are compared to unsigned integers, to
869 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
870 variables that are added to signed variables, to ensure the result is
871 signed and can be negated.
873 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
876 3. Revise the creation of config.h.generic so that all boolean macros are
877 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
878 overriding via -D on the command line possible.
880 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
881 to unsigned int is reported to make a quite noticeable speed difference in
882 a specific Windows environment. Testing on Linux did also appear to show
883 some benefit (and it is clearly not harmful). Also fixed the definition of
884 Xop which should be unsigned.
886 5. Related to (4), changing the definition of the intermediate variable cc
887 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
890 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
893 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
896 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
898 9. Optimizing fast_forward_start_bits in JIT.
900 10. Adding support for callouts in JIT, and fixing some issues revealed
901 during this work. Namely:
903 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
905 (b) Minimum length was not checked before the matching is started.
907 11. The value of capture_last that is passed to callouts was incorrect in some
908 cases when there was a capture on one path that was subsequently abandoned
909 after a backtrack. Also, the capture_last value is now reset after a
910 recursion, since all captures are also reset in this case.
912 12. The interpreter no longer returns the "too many substrings" error in the
913 case when an overflowing capture is in a branch that is subsequently
914 abandoned after a backtrack.
916 13. In the pathological case when an offset vector of size 2 is used, pcretest
917 now prints out the matched string after a yield of 0 or 1.
919 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
920 Only supported by the JIT compiler at the moment.
922 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
924 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
925 offset of the starting point of the matching process, provided the offsets
926 vector is large enough.
928 17. The \A escape now records a lookbehind value of 1, though its execution
929 does not actually inspect the previous character. This is to ensure that,
930 in partial multi-segment matching, at least one character from the old
931 segment is retained when a new segment is processed. Otherwise, if there
932 are no lookbehinds in the pattern, \A might match incorrectly at the start
935 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
937 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
938 gives some modest performance improvement in 8-bit mode.
940 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
941 be expressed in certain programming languages using Universal Character
944 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
945 which points out that "non characters" are not "characters that may not
946 appear in Unicode strings" but rather "characters that are reserved for
947 internal use and have only local meaning".
949 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
950 there was a conditional group that depended on an assertion, if the
951 assertion was false, the callout that immediately followed the alternation
952 in the condition was skipped when pcre_exec() was used for matching.
954 23. Allow an explicit callout to be inserted before an assertion that is the
955 condition for a conditional group, for compatibility with automatic
956 callouts, which always insert a callout at this point.
958 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
959 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
961 25. (*PRUNE) is now supported by the JIT compiler.
963 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
965 27. Fix the case where there are two or more SKIPs with arguments that may be
968 28. (*SKIP) is now supported by the JIT compiler.
970 29. (*THEN) is now supported by the JIT compiler.
972 30. Update RunTest with additional test selector options.
974 31. The way PCRE handles backtracking verbs has been changed in two ways.
976 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
977 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
978 backtracking. In some cases this makes it more Perl-compatible, but Perl's
979 rather obscure rules do not always do the same thing.
981 (2) Previously, backtracking verbs were confined within assertions. This is
982 no longer the case for positive assertions, except for (*ACCEPT). Again,
983 this sometimes improves Perl compatibility, and sometimes does not.
985 32. A number of tests that were in test 2 because Perl did things differently
986 have been moved to test 1, because either Perl or PCRE has changed, and
987 these tests are now compatible.
989 32. Backtracking control verbs are now handled in the same way in JIT and
992 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
993 contained a forward subroutine reference caused a compile error.
995 34. Auto-detect and optimize limited repetitions in JIT.
997 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1000 36. In the interpreter, maximizing pattern repetitions for characters and
1001 character types now use tail recursion, which reduces stack usage.
1003 37. The value of the max lookbehind was not correctly preserved if a compiled
1004 and saved regex was reloaded on a host of different endianness.
1006 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1007 of the compiled pattern block, expand the flags field from 16 to 32 bits
1008 because it was almost full.
1010 39. Try madvise first before posix_madvise.
1012 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1013 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1014 original change locked out all matches of empty strings. This has been
1015 changed so that one match of an empty string per line is recognized.
1016 Subsequent searches on the same line (for colouring or for --only-matching,
1017 for example) do not recognize empty strings.
1019 41. Applied a user patch to fix a number of spelling mistakes in comments.
1021 42. Data lines longer than 65536 caused pcretest to crash.
1023 43. Clarified the data type for length and startoffset arguments for pcre_exec
1024 and pcre_dfa_exec in the function-specific man pages, where they were
1025 explicitly stated to be in bytes, never having been updated. I also added
1026 some clarification to the pcreapi man page.
1028 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1029 a segmentation fault.
1032 Version 8.32 30-November-2012
1033 -----------------------------
1035 1. Improved JIT compiler optimizations for first character search and single
1036 character iterators.
1038 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1039 Patch by Daniel Richard G.
1041 3. Single character iterator optimizations in the JIT compiler.
1043 4. Improved JIT compiler optimizations for character ranges.
1045 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1046 Reported by Giuseppe D'Angelo.
1048 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1049 of a line, was being set incorrectly in cases where .* appeared inside
1050 atomic brackets at the start of a pattern, or where there was a subsequent
1053 7. Improved instruction cache flush for POWER/PowerPC.
1054 Patch by Daniel Richard G.
1056 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
1059 (a) There is now no limit to the number of patterns to be matched.
1061 (b) An error is given if a pattern is too long.
1063 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1066 (d) --exclude-from and --include-from (multiple use) have been added.
1068 (e) Exclusions and inclusions now apply to all files and directories, not
1069 just to those obtained from scanning a directory recursively.
1071 (f) Multiple uses of -f and --file-list are now supported.
1073 (g) In a Windows environment, the default for -d has been changed from
1074 "read" (the GNU grep default) to "skip", because otherwise the presence
1075 of a directory in the file list provokes an error.
1077 (h) The documentation has been revised and clarified in places.
1079 9. Improve the matching speed of capturing brackets.
1081 10. Changed the meaning of \X so that it now matches a Unicode extended
1084 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1085 out POSIX threads when JIT support is configured.
1087 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1089 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1090 structure could go wrong in environments where size_t is not the same size
1093 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1096 15. The EBCDIC support had decayed; later updates to the code had included
1097 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1098 general tidy up of EBCDIC-related issues, and the documentation was also
1099 not quite right. There is now a test that can be run on ASCII systems to
1100 check some of the EBCDIC-related things (but is it not a full test).
1102 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1103 in a small tidy to the code.
1105 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1107 18. If the --only-matching (-o) option in pcregrep is specified multiple
1108 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1109 substrings matched by the 1st and 2nd capturing parentheses. A separating
1110 string can be specified by --om-separator (default empty).
1112 19. Improving the first n character searches.
1114 20. Turn case lists for horizontal and vertical white space into macros so that
1115 they are defined only once.
1117 21. This set of changes together give more compatible Unicode case-folding
1118 behaviour for characters that have more than one other case when UCP
1119 support is available.
1121 (a) The Unicode property table now has offsets into a new table of sets of
1122 three or more characters that are case-equivalent. The MultiStage2.py
1123 script that generates these tables (the pcre_ucd.c file) now scans
1124 CaseFolding.txt instead of UnicodeData.txt for character case
1127 (b) The code for adding characters or ranges of characters to a character
1128 class has been abstracted into a generalized function that also handles
1129 case-independence. In UTF-mode with UCP support, this uses the new data
1130 to handle characters with more than one other case.
1132 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1133 whose other case is greater than 256 are now correctly matched
1134 caselessly. Previously, the high codepoint matched the low one, but not
1137 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1138 of the new class addition function, using character lists defined as
1139 macros alongside the case definitions of 20 above.
1141 (e) Caseless back references now work with characters that have more than
1144 (f) General caseless matching of characters with more than one other case
1147 22. Unicode character properties were updated from Unicode 6.2.0
1149 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1151 24. Add support for 32-bit character strings, and UTF-32
1153 25. Major JIT compiler update (code refactoring and bugfixing).
1154 Experimental Sparc 32 support is added.
1156 26. Applied a modified version of Daniel Richard G's patch to create
1157 pcre.h.generic and config.h.generic by "make" instead of in the
1158 PrepareRelease script.
1160 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1161 pcre_compile.c when checking for a zero character.
1163 28. Introducing a native interface for JIT. Through this interface, the compiled
1164 machine code can be directly executed. The purpose of this interface is to
1165 provide fast pattern matching, so several sanity checks are not performed.
1166 However, feature tests are still performed. The new interface provides
1167 1.4x speedup compared to the old one.
1169 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1170 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1171 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1173 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1174 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1175 the "old" RFC 2279). Instead, it ended up passing a negative length to
1178 31. Add support for GCC's visibility feature to hide internal functions.
1180 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1181 "unknown -C option" after outputting 0 or 1.
1183 33. There is now support for generating a code coverage report for the test
1184 suite in environments where gcc is the compiler and lcov is installed. This
1185 is mainly for the benefit of the developers.
1187 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1188 unaddressable using valgrind annotations, allowing valgrind to detect
1189 invalid memory accesses. This is mainly for the benefit of the developers.
1191 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1193 26. Give configure error if --enable-cpp but no C++ compiler found.
1196 Version 8.31 06-July-2012
1197 -------------------------
1199 1. Fixing a wrong JIT test case and some compiler warnings.
1201 2. Removed a bashism from the RunTest script.
1203 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1204 to unsigned type, result still unsigned" that was given by an MS compiler
1205 on encountering the code "-sizeof(xxx)".
1207 4. Partial matching support is added to the JIT compiler.
1209 5. Fixed several bugs concerned with partial matching of items that consist
1210 of more than one character:
1212 (a) /^(..)\1/ did not partially match "aba" because checking references was
1213 done on an "all or nothing" basis. This also applied to repeated
1216 (b) \R did not give a hard partial match if \r was found at the end of the
1219 (c) \X did not give a hard partial match after matching one or more
1220 characters at the end of the subject.
1222 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1223 a partial match for the string "\r".
1225 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1226 a partial match for a CR character at the end of the subject string.
1228 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1229 running pcretest, the text "(JIT)" added to the output whenever JIT is
1230 actually used to run the match.
1232 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1233 or /S+[+] with a digit between 1 and 7.
1235 8. OP_NOT now supports any UTF character not just single-byte ones.
1237 9. (*MARK) control verb is now supported by the JIT compiler.
1239 10. The command "./RunTest list" lists the available tests without actually
1240 running any of them. (Because I keep forgetting what they all are.)
1242 11. Add PCRE_INFO_MAXLOOKBEHIND.
1244 12. Applied a (slightly modified) user-supplied patch that improves performance
1245 when the heap is used for recursion (compiled with --disable-stack-for-
1246 recursion). Instead of malloc and free for each heap frame each time a
1247 logical recursion happens, frames are retained on a chain and re-used where
1248 possible. This sometimes gives as much as 30% improvement.
1250 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1253 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1255 15. It is now possible to link pcretest with libedit as an alternative to
1258 16. (*COMMIT) control verb is now supported by the JIT compiler.
1260 17. The Unicode data tables have been updated to Unicode 6.1.0.
1262 18. Added --file-list option to pcregrep.
1264 19. Added binary file support to pcregrep, including the -a, --binary-files,
1265 -I, and --text options.
1267 20. The madvise function is renamed for posix_madvise for QNX compatibility
1268 reasons. Fixed by Giuseppe D'Angelo.
1270 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1272 22. Changed the option for creating symbolic links for 16-bit man pages from
1273 -s to -sf so that re-installing does not cause issues.
1275 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1277 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1278 matches in certain environments (the workspace was not being correctly
1279 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1280 some of the workspace data at the beginning of a restart.
1282 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1283 was not doing so when it should - probably a typo introduced by SVN 528
1286 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1287 \w+ when the character tables indicated that \x{c4} was a word character.
1288 There were several related cases, all because the tests for doing a table
1289 lookup were testing for characters less than 127 instead of 255.
1291 27. If a pattern contains capturing parentheses that are not used in a match,
1292 their slots in the ovector are set to -1. For those that are higher than
1293 any matched groups, this happens at the end of processing. In the case when
1294 there were back references that the ovector was too small to contain
1295 (causing temporary malloc'd memory to be used during matching), and the
1296 highest capturing number was not used, memory off the end of the ovector
1297 was incorrectly being set to -1. (It was using the size of the temporary
1298 memory instead of the true size.)
1300 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1301 ovector size, it uses memory at the end of the block that it has got.
1303 29. Check for an overlong MARK name and give an error at compile time. The
1304 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1306 30. JIT compiler update.
1308 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1311 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1313 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1315 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1316 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1319 35. Improve JIT code generation for greedy plus quantifier.
1321 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1322 1 to "aa" instead of to an empty string. The bug affected repeated groups
1323 that could potentially match an empty string.
1325 37. Optimizing single character iterators in JIT.
1327 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1328 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1329 codepoints that are too big for the mode are faulted, and in a UTF mode,
1330 disallowed codepoints are also faulted.
1332 39. If PCRE was compiled with UTF support, in three places in the DFA
1333 matcher there was code that should only have been obeyed in UTF mode, but
1334 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1335 processing when bytes with values greater than 127 were present. In 16-bit
1336 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1337 both cases the values are those that cannot be the first data item in a UTF
1338 character. The three items that might have provoked this were recursions,
1339 possessively repeated groups, and atomic groups.
1341 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1342 and pcregrep, because some OS require shared objects to be explicitly
1343 passed to ld, causing the link step to fail if they are not.
1345 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1346 mode, patterns that started with \h* or \R* might be incorrectly matched.
1349 Version 8.30 04-February-2012
1350 -----------------------------
1352 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1353 name is defined in ctype.h.
1355 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1356 only in quite long subpatterns.
1358 3. Removed the function pcre_info(), which has been obsolete and deprecated
1359 since it was replaced by pcre_fullinfo() in February 2000.
1361 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1362 match a (*MARK), and the match failed at the start of the subject, a
1363 reference to memory before the start of the subject could occur. This bug
1364 was introduced by fix 17 of release 8.21.
1366 5. A reference to an unset group with zero minimum repetition was giving
1367 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1368 /(another)?(\1?)test/ matched against "hello world test". This bug was
1369 introduced in release 8.13.
1371 6. Add support for 16-bit character strings (a large amount of work involving
1372 many changes and refactorings).
1374 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1375 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1378 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1379 rounding is not applied in this particular case).
1381 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1382 if they appear, or are escaped, in patterns.
1384 10. Get rid of a number of -Wunused-but-set-variable warnings.
1386 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1387 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1388 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1389 also returns the mark "x". This bug applied to capturing parentheses,
1390 non-capturing parentheses, and atomic parentheses. It also applied to some
1393 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1394 information out of configure.ac instead of relying on pcre.h.generic, which
1395 is not stored in the repository.
1397 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1400 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1402 15. Applied Graycode's patch to put the top-level frame on the stack rather
1403 than the heap when not using the stack for recursion. This gives a
1404 performance improvement in many cases when recursion is not deep.
1406 16. Experimental code added to "pcretest -C" to output the stack frame size.
1409 Version 8.21 12-Dec-2011
1410 ------------------------
1412 1. Updating the JIT compiler.
1414 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1417 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1418 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1419 calling _pcre_jit_exec. Some extra comments are added.
1421 4. (*MARK) settings inside atomic groups that do not contain any capturing
1422 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1423 was introduced by change 18 for 8.20.
1425 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1428 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1429 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1430 This bug was probably introduced by change 9 of 8.13.
1432 7. While fixing 6 above, I noticed that a number of other items were being
1433 incorrectly rejected as "not fixed length". This arose partly because newer
1434 opcodes had not been added to the fixed-length checking code. I have (a)
1435 corrected the bug and added tests for these items, and (b) arranged for an
1436 error to occur if an unknown opcode is encountered while checking for fixed
1437 length instead of just assuming "not fixed length". The items that were
1438 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1439 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1440 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1442 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1443 being incorrectly compiled and would have given unpredicatble results.
1445 9. A possessively repeated subpattern with minimum repeat count greater than
1446 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1447 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1448 the first (A) could occur when it should not.
1450 10. Add a cast and remove a redundant test from the code.
1452 11. JIT should use pcre_malloc/pcre_free for allocation.
1454 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1455 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1456 is anything other than /usr, -L is still shown).
1458 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1460 14. Perl does not support \N without a following name in a [] class; PCRE now
1461 also gives an error.
1463 15. If a forward reference was repeated with an upper limit of around 2000,
1464 it caused the error "internal error: overran compiling workspace". The
1465 maximum number of forward references (including repeats) was limited by the
1466 internal workspace, and dependent on the LINK_SIZE. The code has been
1467 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1468 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1469 of around 200,000 forward references. While doing this, I also speeded up
1470 the filling in of repeated forward references.
1472 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1473 incorrectly expecting the subject to contain another "a" after the start.
1475 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1476 in the match, the SKIP should be ignored. This was not happening; instead
1477 the SKIP was being treated as NOMATCH. For patterns such as
1478 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1481 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1482 now much more compatible with Perl, in particular in cases where the result
1483 is a non-match for a non-anchored pattern. For example, if
1484 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1485 "m", where previously it did not return a name. A side effect of this
1486 change is that for partial matches, the last encountered mark name is
1487 returned, as for non matches. A number of tests that were previously not
1488 Perl-compatible have been moved into the Perl-compatible test files. The
1489 refactoring has had the pleasing side effect of removing one argument from
1490 the match() function, thus reducing its stack requirements.
1492 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1493 subsequent uses of /S (without +) incorrectly behaved like /S+.
1495 21. Retrieve executable code size support for the JIT compiler and fixing
1498 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1499 not work when the shorter character appeared right at the end of the
1502 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1505 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1506 output it when the /M option is used in pcretest.
1508 25. The CheckMan script was not being included in the distribution. Also, added
1509 an explicit "perl" to run Perl scripts from the PrepareRelease script
1510 because this is reportedly needed in Windows.
1512 26. If study data was being save in a file and studying had not found a set of
1513 "starts with" bytes for the pattern, the data written to the file (though
1514 never used) was taken from uninitialized memory and so caused valgrind to
1517 27. Updated RunTest.bat as provided by Sheri Pierce.
1519 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1521 29. Computation of memory usage for the table of capturing group names was
1522 giving an unnecessarily large value.
1525 Version 8.20 21-Oct-2011
1526 ------------------------
1528 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1529 a POSIX class. After further experiments with Perl, which convinced me that
1530 Perl has bugs and confusions, a closing square bracket is no longer allowed
1531 in a POSIX name. This bug also affected patterns with classes that started
1534 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1535 captured substring, but while checking the failing first alternative,
1536 substring 1 is temporarily captured. If the output vector supplied to
1537 pcre_exec() was not big enough for this capture, the yield of the function
1538 was still zero ("insufficient space for captured substrings"). This cannot
1539 be totally fixed without adding another stack variable, which seems a lot
1540 of expense for a edge case. However, I have improved the situation in cases
1541 such as /(a)(b)x|abc/ matched against "abc", where the return code
1542 indicates that fewer than the maximum number of slots in the ovector have
1545 3. Related to (2) above: when there are more back references in a pattern than
1546 slots in the output vector, pcre_exec() uses temporary memory during
1547 matching, and copies in the captures as far as possible afterwards. It was
1548 using the entire output vector, but this conflicts with the specification
1549 that only 2/3 is used for passing back captured substrings. Now it uses
1550 only the first 2/3, for compatibility. This is, of course, another edge
1553 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1554 main code base, and can be used by building with --enable-jit. When this is
1555 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1556 runtime --no-jit option is given.
1558 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1559 ovector, the return from the function was zero, implying that there were
1560 other matches that did not fit. The correct "exactly full" value is now
1563 6. If a subpattern that was called recursively or as a subroutine contained
1564 (*PRUNE) or any other control that caused it to give a non-standard return,
1565 invalid errors such as "Error -26 (nested recursion at the same subject
1566 position)" or even infinite loops could occur.
1568 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1569 computing the minimum length on reaching *ACCEPT, and so ended up with the
1570 wrong value of 1 rather than 0. Further investigation indicates that
1571 computing a minimum subject length in the presence of *ACCEPT is difficult
1572 (think back references, subroutine calls), and so I have changed the code
1573 so that no minimum is registered for a pattern that contains *ACCEPT.
1575 8. If (*THEN) was present in the first (true) branch of a conditional group,
1576 it was not handled as intended. [But see 16 below.]
1578 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1581 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1582 the first byte in a match must be "a".
1584 11. Change 17 for 8.13 increased the recursion depth for patterns like
1585 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1586 pattern contains any instances of (*THEN). If it does not, the old
1587 optimizations are restored. It would be nice to do this on a per-group
1588 basis, but at the moment that is not feasible.
1590 12. In some environments, the output of pcretest -C is CRLF terminated. This
1591 broke RunTest's code that checks for the link size. A single white space
1592 character after the value is now allowed for.
1594 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1595 For "fr", it uses the Windows-specific input and output files.
1597 14. If (*THEN) appeared in a group that was called recursively or as a
1598 subroutine, it did not work as intended. [But see next item.]
1600 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1601 pattern fragments (but not containing any | characters). If A and B are
1602 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1603 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1604 D. In other words, Perl considers parentheses that do not contain any |
1605 characters to be part of a surrounding alternative, whereas PCRE was
1606 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1607 differently. PCRE now behaves in the same way as Perl, except in the case
1608 of subroutine/recursion calls such as (?1) which have in any case always
1609 been different (but PCRE had them first :-).
1611 16. Related to 15 above: Perl does not treat the | in a conditional group as
1612 creating alternatives. Such a group is treated in the same way as an
1613 ordinary group without any | characters when processing (*THEN). PCRE has
1614 been changed to match Perl's behaviour.
1616 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1617 RunGrepTest script failed.
1619 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1620 inevitable for groups that contain captures, but it can lead to a lot of
1621 stack use in large patterns. The old behaviour has been restored for atomic
1622 groups that do not contain any capturing parentheses.
1624 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1625 suppress the check for a minimum subject length at run time. (If it was
1626 given to pcre_exec() or pcre_dfa_exec() it did work.)
1628 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1629 fail to work when decoding hex characters in data strings in EBCDIC
1632 21. It appears that in at least one Mac OS environment, the isxdigit() function
1633 is implemented as a macro that evaluates to its argument more than once,
1634 contravening the C 90 Standard (I haven't checked a later standard). There
1635 was an instance in pcretest which caused it to go wrong when processing
1636 \x{...} escapes in subject strings. The has been rewritten to avoid using
1637 things like p++ in the argument of isxdigit().
1640 Version 8.13 16-Aug-2011
1641 ------------------------
1643 1. The Unicode data tables have been updated to Unicode 6.0.0.
1645 2. Two minor typos in pcre_internal.h have been fixed.
1647 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1648 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1649 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1651 4. There were a number of related bugs in the code for matching backrefences
1652 caselessly in UTF-8 mode when codes for the characters concerned were
1653 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1654 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1655 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1656 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1657 2-byte code at the end of the subject (it thought there wasn't enough data
1660 5. Comprehensive information about what went wrong is now returned by
1661 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1662 as the output vector has at least 2 elements. The offset of the start of
1663 the failing character and a reason code are placed in the vector.
1665 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1666 now returned is for the first byte of the failing character, instead of the
1667 last byte inspected. This is an incompatible change, but I hope it is small
1668 enough not to be a problem. It makes the returned offset consistent with
1669 pcre_exec() and pcre_dfa_exec().
1671 7. pcretest now gives a text phrase as well as the error number when
1672 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1673 failure, the offset and reason code are output.
1675 8. When \R was used with a maximizing quantifier it failed to skip backwards
1676 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1677 back over a single character (\n). This seems wrong (because it treated the
1678 two characters as a single entity when going forwards), conflicts with the
1679 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1680 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1683 9. Some internal refactoring has changed the processing so that the handling
1684 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1685 time (the PCRE_DOTALL option was changed this way some time ago: version
1686 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1687 which was always a bit of a fudge. It also means that there is one less
1688 argument for the match() function, which reduces its stack requirements
1689 slightly. This change also fixes an incompatibility with Perl: the pattern
1690 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1692 10. More internal refactoring has drastically reduced the number of recursive
1693 calls to match() for possessively repeated groups such as (abc)++ when
1696 11. While implementing 10, a number of bugs in the handling of groups were
1697 discovered and fixed:
1699 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1700 (a|)*(?1) gave a compile-time internal error.
1701 ((a|)+)+ did not notice that the outer group could match an empty string.
1702 (^a|^)+ was not marked as anchored.
1703 (.*a|.*)+ was not marked as matching at start or after a newline.
1705 12. Yet more internal refactoring has removed another argument from the match()
1706 function. Special calls to this function are now indicated by setting a
1707 value in a variable in the "match data" data block.
1709 13. Be more explicit in pcre_study() instead of relying on "default" for
1710 opcodes that mean there is no starting character; this means that when new
1711 ones are added and accidentally left out of pcre_study(), testing should
1714 14. The -s option of pcretest has been documented for ages as being an old
1715 synonym of -m (show memory usage). I have changed it to mean "force study
1716 for every regex", that is, assume /S for every regex. This is similar to -i
1717 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1718 using it. It makes it easier to run collections of tests with and without
1719 study enabled, and thereby test pcre_study() more easily. All the standard
1720 tests are now run with and without -s (but some patterns can be marked as
1721 "never study" - see 20 below).
1723 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1724 restoration of the capturing data to the outer values was not happening
1727 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1728 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1729 pattern had matched an empty string, and so incorrectly returned a no
1732 17. There was optimizing code for the last branch of non-capturing parentheses,
1733 and also for the obeyed branch of a conditional subexpression, which used
1734 tail recursion to cut down on stack usage. Unfortunately, now that there is
1735 the possibility of (*THEN) occurring in these branches, tail recursion is
1736 no longer possible because the return has to be checked for (*THEN). These
1737 two optimizations have therefore been removed. [But see 8.20/11 above.]
1739 18. If a pattern containing \R was studied, it was assumed that \R always
1740 matched two bytes, thus causing the minimum subject length to be
1741 incorrectly computed because \R can also match just one byte.
1743 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1744 was incorrectly computed.
1746 20. If /S is present twice on a test pattern in pcretest input, it now
1747 *disables* studying, thereby overriding the use of -s on the command line
1748 (see 14 above). This is necessary for one or two tests to keep the output
1749 identical in both cases.
1751 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1752 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1754 22. When an atomic group that contained a capturing parenthesis was
1755 successfully matched, but the branch in which it appeared failed, the
1756 capturing was not being forgotten if a higher numbered group was later
1757 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1758 group 1 to "a", when in fact it should be unset. This applied to multi-
1759 branched capturing and non-capturing groups, repeated or not, and also to
1760 positive assertions (capturing in negative assertions does not happen
1761 in PCRE) and also to nested atomic groups.
1763 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1764 subject after a captured substring, to make it easier to tell which of a
1765 number of identical substrings has been captured.
1767 24. The way atomic groups are processed by pcre_exec() has been changed so that
1768 if they are repeated, backtracking one repetition now resets captured
1769 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1770 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1771 "aaa". Previously, it would have been "a". As part of this code
1772 refactoring, the way recursive calls are handled has also been changed.
1774 25. If an assertion condition captured any substrings, they were not passed
1775 back unless some other capturing happened later. For example, if
1776 (?(?=(a))a) was matched against "a", no capturing was returned.
1778 26. When studying a pattern that contained subroutine calls or assertions,
1779 the code for finding the minimum length of a possible match was handling
1780 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1781 group 1 called group 2 while simultaneously a separate group 2 called group
1782 1). A stack overflow occurred in this case. I have fixed this by limiting
1783 the recursion depth to 10.
1785 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1786 Fortmann. This supports explicit test numbers on the command line, and has
1787 argument validation and error reporting.
1789 28. An instance of \X with an unlimited repeat could fail if at any point the
1790 first character it looked at was a mark character.
1792 29. Some minor code refactoring concerning Unicode properties and scripts
1793 should reduce the stack requirement of match() slightly.
1795 30. Added the '=' option to pcretest to check the setting of unused capturing
1796 slots at the end of the pattern, which are documented as being -1, but are
1797 not included in the return count.
1799 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1800 compiled something random. Now it gives a compile-time error (as does
1803 32. A *MARK encountered during the processing of a positive assertion is now
1804 recorded and passed back (compatible with Perl).
1806 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1807 had alternative anchored branches, the search for a second match in a line
1808 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1809 matched the line "0102" twice. The same bug affected patterns that started
1810 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1813 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1814 does, and because of capturing effects, quantifying parenthesized
1815 assertions may at times be useful. Quantifiers are now allowed for
1816 parenthesized assertions.
1818 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1820 36. \g was being checked for fancy things in a character class, when it should
1821 just be a literal "g".
1823 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1824 appearance of a nested POSIX class supersedes an apparent external class.
1825 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1826 unescaped square brackets may also appear as part of class names. For
1827 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1828 more like Perl. (But see 8.20/1 above.)
1830 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1831 was because it thought it was \N{name}, which is not supported).
1833 39. Add minix to OS list not supporting the -S option in pcretest.
1835 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1836 cannot analyze patterns in sufficient detail to catch mutual recursions
1837 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1838 subgroup is called recursively as a subpattern for a second time at the
1839 same position in the subject string. In previous releases this might have
1840 been caught by the recursion limit, or it might have run out of stack.
1842 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1843 happen only once. PCRE was, however incorrectly giving a compile time error
1844 "recursive call could loop indefinitely" because it cannot analyze the
1845 pattern in sufficient detail. The compile time test no longer happens when
1846 PCRE is compiling a conditional subpattern, but actual runaway loops are
1847 now caught at runtime (see 40 above).
1849 42. It seems that Perl allows any characters other than a closing parenthesis
1850 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1851 has been changed to be the same.
1853 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1854 as not to get warnings when autogen.sh is called. Also changed
1855 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1857 44. To help people who use pcregrep to scan files containing exceedingly long
1858 lines, the following changes have been made:
1860 (a) The default value of the buffer size parameter has been increased from
1861 8K to 20K. (The actual buffer used is three times this size.)
1863 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1866 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1867 to be set at run time.
1869 (d) Numerical values in pcregrep options can be followed by K or M, for
1870 example --buffer-size=50K.
1872 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1873 given and the return code is set to 2.
1875 45. Add a pointer to the latest mark to the callout data block.
1877 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1878 partial match of an empty string instead of no match. This was specific to
1881 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1882 complete match instead of a partial match. This bug was dependent on both
1883 the PCRE_UTF8 and PCRE_DOTALL options being set.
1885 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1886 starting byte set, because \b was not being ignored.
1889 Version 8.12 15-Jan-2011
1890 ------------------------
1892 1. Fixed some typos in the markup of the man pages, and wrote a script that
1893 checks for such things as part of the documentation building process.
1895 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1896 --match-limit and --recursion-limit options (added for 8.11). In
1897 particular, this made one of the standard tests fail. (The integer value
1898 went into the wrong half of a long int.)
1900 3. If the --colour option was given to pcregrep with -v (invert match), it
1901 did strange things, either producing crazy output, or crashing. It should,
1902 of course, ignore a request for colour when reporting lines that do not
1905 4. Another pcregrep bug caused similar problems if --colour was specified with
1906 -M (multiline) and the pattern match finished with a line ending.
1908 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1909 matched in multiline mode, the following line was shown as part of the
1910 match. This seems wrong, so I have changed it.
1912 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1913 the check for further matches in the same line (so they could be coloured)
1914 to overrun the end of the current line. If another match was found, it was
1915 incorrectly shown (and then shown again when found in the next line).
1917 7. If pcregrep was compiled under Windows, there was a reference to the
1918 function pcregrep_exit() before it was defined. I am assuming this was
1919 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1920 reported by a user. I've moved the definition above the reference.
1923 Version 8.11 10-Dec-2010
1924 ------------------------
1926 1. (*THEN) was not working properly if there were untried alternatives prior
1927 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1928 backtracked to try for "b" instead of moving to the next alternative branch
1929 at the same level (in this case, to look for "c"). The Perl documentation
1930 is clear that when (*THEN) is backtracked onto, it goes to the "next
1931 alternative in the innermost enclosing group".
1933 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1934 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1935 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1936 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1939 3. If \s appeared in a character class, it removed the VT character from
1940 the class, even if it had been included by some previous item, for example
1941 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1942 of \s, but is part of the POSIX "space" class.)
1944 4. A partial match never returns an empty string (because you can always
1945 match an empty string at the end of the subject); however the checking for
1946 an empty string was starting at the "start of match" point. This has been
1947 changed to the "earliest inspected character" point, because the returned
1948 data for a partial match starts at this character. This means that, for
1949 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1950 (previously it gave "no match").
1952 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1953 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1954 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1955 has an implication that the given string is incomplete (because a partial
1956 match is preferred over a full match). For this reason, these items now
1957 give a partial match in this situation. [Aside: previously, the one case
1958 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1959 match rather than a full match, which was wrong by the old rules, but is
1962 6. There was a bug in the handling of #-introduced comments, recognized when
1963 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1964 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1965 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1966 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1967 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1968 places in pcre_compile().
1970 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1971 comments when looking ahead for named forward references to subpatterns,
1972 the only newline sequence it recognized was NL. It now handles newlines
1973 according to the set newline convention.
1975 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1976 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1977 cater for a lack of strerror(). These oversights have been fixed.
1979 9. Added --match-limit and --recursion-limit to pcregrep.
1981 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1983 11. When the -o option was used, pcregrep was setting a return code of 1, even
1984 when matches were found, and --line-buffered was not being honoured.
1986 12. Added an optional parentheses number to the -o and --only-matching options
1989 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1990 can match an empty string. The code to do it in pcretest and pcredemo
1993 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1994 only one byte after an empty string match just before CRLF (this case
1995 just got forgotten; "any" and "anycrlf" were OK).
1997 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1998 mode when an empty string match preceded an ASCII character followed by
1999 a non-ASCII character. (The code for advancing by one character rather
2000 than one byte was nonsense.)
2002 (c) The pcredemo.c sample program did not have any code at all to handle
2003 the cases when CRLF is a valid newline sequence.
2005 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2006 as a starting offset was within the subject string. There is now a new
2007 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2008 negative or greater than the length of the string. In order to test this,
2009 pcretest is extended to allow the setting of negative starting offsets.
2011 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2012 starting offset points to the beginning of a UTF-8 character was
2013 unnecessarily clumsy. I tidied it up.
2015 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2016 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2018 17. Nobody had reported that the --include_dir option, which was added in
2019 release 7.7 should have been called --include-dir (hyphen, not underscore)
2020 for compatibility with GNU grep. I have changed it to --include-dir, but
2021 left --include_dir as an undocumented synonym, and the same for
2022 --exclude-dir, though that is not available in GNU grep, at least as of
2025 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2026 characters from a string of bytes) have been redefined so as not to use
2027 loops, in order to improve performance in some environments. At the same
2028 time, I abstracted some of the common code into auxiliary macros to save
2029 repetition (this should not affect the compiled code).
2031 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2032 compile-time error is now given if \c is not followed by an ASCII
2033 character, that is, a byte less than 128. (In EBCDIC mode, the code is
2034 different, and any byte value is allowed.)
2036 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2037 START_OPTIMIZE option, which is now allowed at compile time - but just
2038 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2039 to pcregrep and other applications that have no direct access to PCRE
2040 options. The new /Y option in pcretest sets this option when calling
2043 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2044 back references. Groups containing recursive back references were forced to
2045 be atomic by that change, but in the case of named groups, the amount of
2046 memory required was incorrectly computed, leading to "Failed: internal
2047 error: code overflow". This has been fixed.
2049 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2050 pcretest.c, to avoid build problems in some Borland environments.
2053 Version 8.10 25-Jun-2010
2054 ------------------------
2056 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2059 2. (*ACCEPT) was not working when inside an atomic group.
2061 3. Inside a character class, \B is treated as a literal by default, but
2062 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2063 causes the error). The code is unchanged, but I tidied the documentation.
2065 4. Inside a character class, PCRE always treated \R and \X as literals,
2066 whereas Perl faults them if its -w option is set. I have changed PCRE so
2067 that it faults them when PCRE_EXTRA is set.
2069 5. Added support for \N, which always matches any character other than
2070 newline. (It is the same as "." when PCRE_DOTALL is not set.)
2072 6. When compiling pcregrep with newer versions of gcc which may have
2073 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2074 declared with attribute warn_unused_result" were given. Just casting the
2075 result to (void) does not stop the warnings; a more elaborate fudge is
2076 needed. I've used a macro to implement this.
2078 7. Minor change to pcretest.c to avoid a compiler warning.
2080 8. Added four artifical Unicode properties to help with an option to make
2081 \s etc use properties (see next item). The new properties are: Xan
2082 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2084 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2085 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2086 this option. Modified pcretest to add /W to test this facility. Added
2087 REG_UCP to make it available via the POSIX interface.
2089 10. Added --line-buffered to pcregrep.
2091 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2092 studied, and the match started with a letter with a code point greater than
2093 127 whose first byte was different to the first byte of the other case of
2094 the letter, the other case of this starting letter was not recognized
2097 12. If a pattern that was studied started with a repeated Unicode property
2098 test, for example, \p{Nd}+, there was the theoretical possibility of
2099 setting up an incorrect bitmap of starting bytes, but fortunately it could
2100 not have actually happened in practice until change 8 above was made (it
2101 added property types that matched character-matching opcodes).
2103 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2104 possible starting bytes for non-anchored patterns.
2106 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2107 \R, and also a number of cases that involve Unicode properties, both
2108 explicit and implicit when PCRE_UCP is set.
2110 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2111 input, it could crash or give wrong results if characters with values
2112 greater than 0xc0 were present in the subject string. (Detail: it assumed
2113 UTF-8 input when processing these items.)
2115 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2116 size_t is 64-bit (#991).
2118 17. Added a check for running out of memory when PCRE is compiled with
2119 --disable-stack-for-recursion (#990).
2121 18. If the last data line in a file for pcretest does not have a newline on
2122 the end, a newline was missing in the output.
2124 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2125 less than 128) in its various bitmaps. However, there is a facility for
2126 generating tables according to the current locale when PCRE is compiled. It
2127 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2128 characters, are recognized by isspace() and therefore were getting set in
2129 these tables, and indeed these tables seem to approximate to ISO 8859. This
2130 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2131 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2132 which of course cannot start UTF-8 characters. I have changed the code so
2133 that only real ASCII characters (less than 128) and the correct starting
2134 bytes for UTF-8 encodings are set for characters greater than 127 when in
2135 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2138 20. Added the /T option to pcretest so as to be able to run tests with non-
2139 standard character tables, thus making it possible to include the tests
2140 used for 19 above in the standard set of tests.
2142 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2143 reference to a subpattern the other side of a comment that contains an
2144 opening parenthesis caused either an internal compiling error, or a
2145 reference to the wrong subpattern.
2148 Version 8.02 19-Mar-2010
2149 ------------------------
2151 1. The Unicode data tables have been updated to Unicode 5.2.0.
2153 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2156 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2157 original author of that file, following a query about its status.
2159 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2160 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2162 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2163 quantifier applied to a forward-referencing subroutine call, could compile
2164 incorrect code or give the error "internal error: previously-checked
2165 referenced subpattern not found".
2167 6. Both MS Visual Studio and Symbian OS have problems with initializing
2168 variables to point to external functions. For these systems, therefore,
2169 pcre_malloc etc. are now initialized to local functions that call the
2170 relevant global functions.
2172 7. There were two entries missing in the vectors called coptable and poptable
2173 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2174 I've fixed the data, and added a kludgy way of testing at compile time that
2175 the lengths are correct (equal to the number of opcodes).
2177 8. Following on from 7, I added a similar kludge to check the length of the
2178 eint vector in pcreposix.c.
2180 9. Error texts for pcre_compile() are held as one long string to avoid too
2181 much relocation at load time. To find a text, the string is searched,
2182 counting zeros. There was no check for running off the end of the string,
2183 which could happen if a new error number was added without updating the
2186 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2188 11. \K was not working if it appeared in an atomic group or in a group that
2189 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2190 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2191 the assertion is positive, but not if it is negative.
2193 12. Change 11 fortuitously reduced the size of the stack frame used in the
2194 "match()" function of pcre_exec.c by one pointer. Forthcoming
2195 implementation of support for (*MARK) will need an extra pointer on the
2196 stack; I have reserved it now, so that the stack frame size does not
2199 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2200 item in branch that calls a recursion is a subroutine call - as in the
2201 second branch in the above example - was incorrectly given the compile-
2202 time error "recursive call could loop indefinitely" because pcre_compile()
2203 was not correctly checking the subroutine for matching a non-empty string.
2205 14. The checks for overrunning compiling workspace could trigger after an
2206 overrun had occurred. This is a "should never occur" error, but it can be
2207 triggered by pathological patterns such as hundreds of nested parentheses.
2208 The checks now trigger 100 bytes before the end of the workspace.
2210 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2213 Version 8.01 19-Jan-2010
2214 ------------------------
2216 1. If a pattern contained a conditional subpattern with only one branch (in
2217 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2218 computed the wrong minimum data length (which is of course zero for such
2219 subpatterns). This could cause incorrect "no match" results.
2221 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2222 the pattern is reset in the first branch, pcre_compile() failed with
2223 "internal error: code overflow at offset...". This happened only when
2224 the reset was to the original external option setting. (An optimization
2225 abstracts leading options settings into an external setting, which was the
2228 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2229 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2230 assertion pattern did not match (meaning that the assertion was true), it
2231 was incorrectly treated as false if the SKIP had been reached during the
2232 matching. This also applied to assertions used as conditions.
2234 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2235 assertion subpattern, including such a pattern used as a condition,
2236 unpredictable results occurred, instead of the error return
2237 PCRE_ERROR_DFA_UITEM.
2239 5. The C++ GlobalReplace function was not working like Perl for the special
2240 situation when an empty string is matched. It now does the fancy magic
2241 stuff that is necessary.
2243 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2244 removed. (These were left over from very, very early versions of PCRE.)
2246 7. Some cosmetic changes to the code to make life easier when compiling it
2247 as part of something else:
2249 (a) Change DEBUG to PCRE_DEBUG.
2251 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2252 called "current" as "current_branch", to prevent a collision with the
2253 Linux macro when compiled as a kernel module.
2255 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2256 prevent a collision with the Linux macro when compiled as a kernel
2259 8. In pcre_compile() there are some checks for integer overflows that used to
2260 cast potentially large values to (double). This has been changed to that
2261 when building, a check for int64_t is made, and if it is found, it is used
2262 instead, thus avoiding the use of floating point arithmetic. (There is no
2263 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2266 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2267 2005 (difference between two addresses compared to an unsigned value).
2269 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2270 custom one, because of the following reported problem in Windows:
2272 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2274 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2275 therefore missing the function definition.
2276 - The compiler thus generates a "C" signature for the test function.
2277 - The linker fails to find the "C" function.
2278 - PCRE fails to configure if asked to do so against libbz2.
2280 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2281 messages were output:
2283 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2284 rerunning libtoolize, to keep the correct libtool macros in-tree.
2285 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2287 I have done both of these things.
2289 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2290 most of the time, it *can* run out if it is given a pattern that contains a
2291 runaway infinite recursion. I updated the discussion in the pcrestack man
2294 13. Now that we have gone to the x.xx style of version numbers, the minor
2295 version may start with zero. Using 08 or 09 is a bad idea because users
2296 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2297 interpreted as invalid octal numbers. I've updated the previous comment in
2298 configure.ac, and also added a check that gives an error if 08 or 09 are
2301 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2302 causing partial matching to fail when the end of the subject matched \W
2303 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2305 15. There were some discrepancies between the declarations in pcre_internal.h
2306 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2307 their definitions. The declarations used "const uschar *" and the
2308 definitions used USPTR. Even though USPTR is normally defined as "const
2309 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2310 reported that: "This difference in casting confuses some C++ compilers, for
2311 example, SunCC recognizes above declarations as different functions and
2312 generates broken code for hbpcre." I have changed the declarations to use
2315 16. GNU libtool is named differently on some systems. The autogen.sh script now
2316 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2319 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2320 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2321 comment: "Figure out how to create a longlong from a string: strtoll and
2322 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2323 instance, but it only takes 2 args instead of 3!"
2325 18. A subtle bug concerned with back references has been fixed by a change of
2326 specification, with a corresponding code fix. A pattern such as
2327 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2328 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2329 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2330 same bug. Such groups have to be quantified to be useful, or contained
2331 inside another quantified group. (If there's no repetition, the reference
2332 can never match.) The problem arises because, having left the group and
2333 moved on to the rest of the pattern, a later failure that backtracks into
2334 the group uses the captured value from the final iteration of the group
2335 rather than the correct earlier one. I have fixed this in PCRE by forcing
2336 any group that contains a reference to itself to be an atomic group; that
2337 is, there cannot be any backtracking into it once it has completed. This is
2338 similar to recursive and subroutine calls.
2341 Version 8.00 19-Oct-09
2342 ----------------------
2344 1. The table for translating pcre_compile() error codes into POSIX error codes
2345 was out-of-date, and there was no check on the pcre_compile() error code
2346 being within the table. This could lead to an OK return being given in
2349 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2350 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2351 in a Windows environment.
2353 3. The pcregrep --count option prints the count for each file even when it is
2354 zero, as does GNU grep. However, pcregrep was also printing all files when
2355 --files-with-matches was added. Now, when both options are given, it prints
2356 counts only for those files that have at least one match. (GNU grep just
2357 prints the file name in this circumstance, but including the count seems
2358 more useful - otherwise, why use --count?) Also ensured that the
2359 combination -clh just lists non-zero counts, with no names.
2361 4. The long form of the pcregrep -F option was incorrectly implemented as
2362 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2363 but it seems right to fix it, and I didn't think it was worth preserving
2366 5. The command line items --regex=pattern and --regexp=pattern were not
2367 recognized by pcregrep, which required --regex pattern or --regexp pattern
2368 (with a space rather than an '='). The man page documented the '=' forms,
2369 which are compatible with GNU grep; these now work.
2371 6. No libpcreposix.pc file was created for pkg-config; there was just
2372 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2374 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2375 when UCP support is not needed, by modifying the Python script that
2376 generates it from Unicode data files. This should not matter if the module
2377 is correctly used as a library, but I received one complaint about 50K of
2378 unwanted data. My guess is that the person linked everything into his
2379 program rather than using a library. Anyway, it does no harm.
2381 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2382 was a minimum greater than 1 for a wide character in a possessive
2383 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2384 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2385 character. Chaos in the form of incorrect output or a compiling loop could
2388 9. The restrictions on what a pattern can contain when partial matching is
2389 requested for pcre_exec() have been removed. All patterns can now be
2390 partially matched by this function. In addition, if there are at least two
2391 slots in the offset vector, the offset of the earliest inspected character
2392 for the match and the offset of the end of the subject are set in them when
2393 PCRE_ERROR_PARTIAL is returned.
2395 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2396 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2397 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2398 and may be more useful for multi-segment matching.
2400 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2401 used to be given if ever the end of the subject was reached; now it is
2402 given only if matching could not proceed because another character was
2403 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2404 string "Z", which now yields "no match" instead of "partial match". In the
2405 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2406 final character ended with (*FAIL).
2408 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2409 if the pattern had a "must contain" character that was already found in the
2410 earlier partial match, unless partial matching was again requested. For
2411 example, with the pattern /dog.(body)?/, the "must contain" character is
2412 "g". If the first part-match was for the string "dog", restarting with
2413 "sbody" failed. This bug has been fixed.
2415 13. The string returned by pcre_dfa_exec() after a partial match has been
2416 changed so that it starts at the first inspected character rather than the
2417 first character of the match. This makes a difference only if the pattern
2418 starts with a lookbehind assertion or \b or \B (\K is not supported by
2419 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2420 matching functions compatible, and I think it's the right thing to do.
2422 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2423 so that the demonstration program is easily available in environments where
2424 PCRE has not been installed from source.
2426 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2427 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2430 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2431 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2432 is not the first non-POSIX option to be added. Clearly some people find
2433 these options useful.
2435 17. If a caller to the POSIX matching function regexec() passes a non-zero
2436 value for nmatch with a NULL value for pmatch, the value of
2437 nmatch is forced to zero.
2439 18. RunGrepTest did not have a test for the availability of the -u option of
2440 the diff command, as RunTest does. It now checks in the same way as
2441 RunTest, and also checks for the -b option.
2443 19. If an odd number of negated classes containing just a single character
2444 interposed, within parentheses, between a forward reference to a named
2445 subpattern and the definition of the subpattern, compilation crashed with
2446 an internal error, complaining that it could not find the referenced
2447 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2448 [The bug was that it was starting one character too far in when skipping
2449 over the character class, thus treating the ] as data rather than
2450 terminating the class. This meant it could skip too much.]
2452 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2453 /g option in pcretest when the pattern contains \K, which makes it possible
2454 to have an empty string match not at the start, even when the pattern is
2455 anchored. Updated pcretest and pcredemo to use this option.
2457 21. If the maximum number of capturing subpatterns in a recursion was greater
2458 than the maximum at the outer level, the higher number was returned, but
2459 with unset values at the outer level. The correct (outer level) value is
2462 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2463 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2464 make it do so. The string so far is captured, making this feature
2465 compatible with Perl.
2467 23. The tests have been re-organized, adding tests 11 and 12, to make it
2468 possible to check the Perl 5.10 features against Perl 5.10.
2470 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2471 pattern matches a fixed length string. PCRE did not allow this; now it
2472 does. Neither allows recursion.
2474 25. I finally figured out how to implement a request to provide the minimum
2475 length of subject string that was needed in order to match a given pattern.
2476 (It was back references and recursion that I had previously got hung up
2477 on.) This code has now been added to pcre_study(); it finds a lower bound
2478 to the length of subject needed. It is not necessarily the greatest lower
2479 bound, but using it to avoid searching strings that are too short does give
2480 some useful speed-ups. The value is available to calling programs via
2483 26. While implementing 25, I discovered to my embarrassment that pcretest had
2484 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2485 study optimizations had never been tested with that matching function.
2486 Oops. What is worse, even when it was passed study data, there was a bug in
2487 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2488 were also very few tests of studied patterns with pcre_dfa_exec().
2490 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2491 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2492 on the other side of the coin, they are no longer allowed to have different
2493 names, because these cannot be distinguished in PCRE, and this has caused
2494 confusion. (This is a difference from Perl.)
2496 28. When duplicate subpattern names are present (necessarily with different
2497 numbers, as required by 27 above), and a test is made by name in a
2498 conditional pattern, either for a subpattern having been matched, or for
2499 recursion in such a pattern, all the associated numbered subpatterns are
2500 tested, and the overall condition is true if the condition is true for any
2501 one of them. This is the way Perl works, and is also more like the way
2502 testing by number works.
2505 Version 7.9 11-Apr-09
2506 ---------------------
2508 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2509 (pcretest), all targets were linked against these libraries. This included
2510 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2511 libraries. This caused unwanted dependencies to be created. This problem
2512 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2513 pcretest is linked with readline.
2515 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2516 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2517 moved outside it again, because FALSE and TRUE are already defined in AIX,
2520 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2521 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2523 4. The pcregrep documentation said spaces were inserted as well as colons (or
2524 hyphens) following file names and line numbers when outputting matching
2525 lines. This is not true; no spaces are inserted. I have also clarified the
2526 wording for the --colour (or --color) option.
2528 5. In pcregrep, when --colour was used with -o, the list of matching strings
2529 was not coloured; this is different to GNU grep, so I have changed it to be
2532 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2533 each matching line was coloured. Now it goes on to look for further matches
2534 of any of the test patterns, which is the same behaviour as GNU grep.
2536 7. A pattern that could match an empty string could cause pcregrep to loop; it
2537 doesn't make sense to accept an empty string match in pcregrep, so I have
2538 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2539 seems to be how GNU grep behaves. [But see later change 40 for release
2542 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2543 start or after a newline", because the conditional assertion was not being
2544 correctly handled. The rule now is that both the assertion and what follows
2545 in the first alternative must satisfy the test.
2547 9. If auto-callout was enabled in a pattern with a conditional group whose
2548 condition was an assertion, PCRE could crash during matching, both with
2549 pcre_exec() and pcre_dfa_exec().
2551 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2554 11. Unicode property support in character classes was not working for
2555 characters (bytes) greater than 127 when not in UTF-8 mode.
2557 12. Added the -M command line option to pcretest.
2559 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2561 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2563 16. Added comments and documentation about mis-use of no_arg in the C++
2566 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2567 from Martin Jerabek that uses macro names for all relevant character and
2570 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2571 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2572 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2573 these, but not everybody uses configure.
2575 19. A conditional group that had only one branch was not being correctly
2576 recognized as an item that could match an empty string. This meant that an
2577 enclosing group might also not be so recognized, causing infinite looping
2578 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2579 with the subject "ab", where knowledge that the repeated group can match
2580 nothing is needed in order to break the loop.
2582 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2583 exec(), but without supplying a callout function, matching went wrong.
2585 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2586 leak if the size of the offset vector was greater than 30. When the vector
2587 is smaller, the saved offsets during recursion go onto a local stack
2588 vector, but for larger vectors malloc() is used. It was failing to free
2589 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2592 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2593 heapframe that is used only when UTF-8 support is enabled. This caused no
2594 problem, but was untidy.
2596 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2597 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2598 included within another project.
2600 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2601 slightly modified by me:
2603 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2604 not building pcregrep.
2606 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2607 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2609 25. Forward references, both numeric and by name, in patterns that made use of
2610 duplicate group numbers, could behave incorrectly or give incorrect errors,
2611 because when scanning forward to find the reference group, PCRE was not
2612 taking into account the duplicate group numbers. A pattern such as
2613 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2615 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2616 the feature of a custom pointer more persuasive (as requested by a user).
2618 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2619 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2622 28. Added support for (*UTF8) at the start of a pattern.
2624 29. Arrange for flags added by the "release type" setting in CMake to be shown
2625 in the configuration summary.
2628 Version 7.8 05-Sep-08
2629 ---------------------
2631 1. Replaced UCP searching code with optimized version as implemented for Ad
2632 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2633 stage table and inline lookup instead of a function, giving speed ups of 2
2634 to 5 times on some simple patterns that I tested. Permission was given to
2635 distribute the MultiStage2.py script that generates the tables (it's not in
2636 the tarball, but is in the Subversion repository).
2638 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2641 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2642 a group with a zero qualifier. The result of the study could be incorrect,
2643 or the function might crash, depending on the pattern.
2645 4. Caseless matching was not working for non-ASCII characters in back
2646 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2647 It now works when Unicode Property Support is available.
2649 5. In pcretest, an escape such as \x{de} in the data was always generating
2650 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2651 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2654 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2656 7. Added two (int) casts to pcregrep when printing the difference of two
2657 pointers, in case they are 64-bit values.
2659 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2662 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2663 and a #define of that name to empty if it is not externally set. This is to
2664 allow users of MSVC to set it if necessary.
2666 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2667 the convenience functions in the pcre_get.c source file.
2669 11. An option change at the start of a pattern that had top-level alternatives
2670 could cause overwriting and/or a crash. This command provoked a crash in
2673 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2675 This potential security problem was recorded as CVE-2008-2371.
2677 12. For a pattern where the match had to start at the beginning or immediately
2678 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2679 pcre_dfa_exec() could read past the end of the passed subject if there was
2680 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2681 pcretest so that it places the subject at the end of its malloc-ed buffer.
2683 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2684 exec() might read past the end of the data buffer in UTF-8 mode.
2686 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2687 the data contained the byte 0x85 as part of a UTF-8 character within its
2688 first line. This applied both to normal and DFA matching.
2690 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2691 /^[^d]*?$/8 failed to match "abc".
2693 16. Added a missing copyright notice to pcrecpp_internal.h.
2695 17. Make it more clear in the documentation that values returned from
2696 pcre_exec() in ovector are byte offsets, not character counts.
2698 18. Tidied a few places to stop certain compilers from issuing warnings.
2700 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2701 supplied by Stefan Weber. I made a further small update for 7.8 because
2702 there is a change of source arrangements: the pcre_searchfuncs.c module is
2703 replaced by pcre_ucd.c.
2706 Version 7.7 07-May-08
2707 ---------------------
2709 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2710 a string to a long long, pretend we don't even have a long long." This is
2711 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2713 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2714 pre-7.6 versions, which defined a global no_arg variable instead of putting
2715 it in the RE class. (See also #8 below.)
2717 3. Remove a line of dead code, identified by coverity and reported by Nuno
2720 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2722 (1) The include/exclude patterns were being applied to the whole pathnames
2723 of files, instead of just to the final components.
2725 (2) If there was more than one level of directory, the subdirectories were
2726 skipped unless they satisfied the include/exclude conditions. This is
2727 inconsistent with GNU grep (and could even be seen as contrary to the
2728 pcregrep specification - which I improved to make it absolutely clear).
2729 The action now is always to scan all levels of directory, and just
2730 apply the include/exclude patterns to regular files.
2732 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2733 --exclude_dir in the tests to avoid scanning .svn directories.
2735 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2736 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2737 doesn't support NULs in patterns.
2739 7. Added some missing "const"s to declarations of static tables in
2740 pcre_compile.c and pcre_dfa_exec.c.
2742 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2743 caused by fix #2 above. (Subsequently also a second patch to fix the
2744 first patch. And a third patch - this was a messy problem.)
2746 9. Applied Craig's patch to remove the use of push_back().
2748 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2749 matching function regexec().
2751 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2752 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2753 references. PCRE supports relative numbers with this syntax (I don't think
2756 12. Previously, a group with a zero repeat such as (...){0} was completely
2757 omitted from the compiled regex. However, this means that if the group
2758 was called as a subroutine from elsewhere in the pattern, things went wrong
2759 (an internal error was given). Such groups are now left in the compiled
2760 pattern, with a new opcode that causes them to be skipped at execution
2763 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2764 to the way PCRE behaves:
2766 (a) A lone ] character is dis-allowed (Perl treats it as data).
2768 (b) A back reference to an unmatched subpattern matches an empty string
2769 (Perl fails the current match path).
2771 (c) A data ] in a character class must be notated as \] because if the
2772 first data character in a class is ], it defines an empty class. (In
2773 Perl it is not possible to have an empty class.) The empty class []
2774 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2775 The negative empty class [^] matches any one character, independently
2776 of the DOTALL setting.
2778 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2779 non-existent subpattern following a character class starting with ']' and
2780 containing () gave an internal compiling error instead of "reference to
2781 non-existent subpattern". Fortunately, when the pattern did exist, the
2782 compiled code was correct. (When scanning forwards to check for the
2783 existence of the subpattern, it was treating the data ']' as terminating
2784 the class, so got the count wrong. When actually compiling, the reference
2785 was subsequently set up correctly.)
2787 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2788 it was being rejected as not supported by pcre_dfa_exec(), even though
2789 other assertions are supported. I have made pcre_dfa_exec() support
2792 16. The implementation of 13c above involved the invention of a new opcode,
2793 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2794 cannot be changed at match time, I realized I could make a small
2795 improvement to matching performance by compiling OP_ALLANY instead of
2796 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2799 17. Compiling pcretest on Windows with readline support failed without the
2800 following two fixes: (1) Make the unistd.h include conditional on
2801 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2803 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2804 ncurses library to be included for pcretest when ReadLine support is
2805 requested, but also to allow for it to be overridden. This patch came from
2808 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2809 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2810 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2814 Version 7.6 28-Jan-08
2815 ---------------------
2817 1. A character class containing a very large number of characters with
2818 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2821 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2822 HAVE_LONG_LONG is not defined.
2824 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2825 bring them up to date and include new features. This patch includes:
2827 - Fixed PH's badly added libz and libbz2 support.
2828 - Fixed a problem with static linking.
2829 - Added pcredemo. [But later removed - see 7 below.]
2830 - Fixed dftables problem and added an option.
2831 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2833 - Added readline support for pcretest.
2834 - Added an listing of the option settings after cmake has run.
2836 4. A user submitted a patch to Makefile that makes it easy to create
2837 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2838 Makefile.am that cause it to include this special target, without
2839 affecting anything else. Note that the same mingw target plus all
2840 the other distribution libraries and programs are now supported
2841 when configuring with CMake (see 6 below) instead of with
2844 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2845 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2846 exported in the Windows port". It has not yet been confirmed that the patch
2847 solves the problem, but it does no harm.
2849 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2850 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2851 with CMake, and also correct the comment about stack recursion.
2853 7. Remove the automatic building of pcredemo from the ./configure system and
2854 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2855 of a program that users should build themselves after PCRE is installed, so
2856 building it automatically is not really right. What is more, it gave
2857 trouble in some build environments.
2859 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2862 Version 7.5 10-Jan-08
2863 ---------------------
2865 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2866 values in parens when parsing an RE using the C++ wrapper."
2868 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2869 Characters greater than 255 were excluded from the class instead of being
2872 3. The same bug as (2) above applied to negated POSIX classes such as
2875 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2876 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2879 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2880 first named subpattern) and a construct such as (?&a) would reference the
2881 first named subpattern whose name started with "a" (in other words, the
2882 length check was missing). Both these problems are fixed. "Subpattern name
2883 expected" is now given for (?&) (a zero-length name), and this patch also
2884 makes it give the same error for \k'' (previously it complained that that
2885 was a reference to a non-existent subpattern).
2887 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2888 this is right because (?- can be followed by option settings as well as by
2889 digits. I have, however, made the messages clearer.
2891 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2892 than the number used in the conditional) now cause a compile-time error.
2893 This is actually not compatible with Perl, which accepts such patterns, but
2894 treats the conditional as always being FALSE (as PCRE used to), but it
2895 seems to me that giving a diagnostic is better.
2897 8. Change "alphameric" to the more common word "alphanumeric" in comments
2900 9. Fix two occurrences of "backslash" in comments that should have been
2903 10. Remove two redundant lines of code that can never be obeyed (their function
2904 was moved elsewhere).
2906 11. The program that makes PCRE's Unicode character property table had a bug
2907 which caused it to generate incorrect table entries for sequences of
2908 characters that have the same character type, but are in different scripts.
2909 It amalgamated them into a single range, with the script of the first of
2910 them. In other words, some characters were in the wrong script. There were
2911 thirteen such cases, affecting characters in the following ranges:
2927 12. The -o option (show only the matching part of a line) for pcregrep was not
2928 compatible with GNU grep in that, if there was more than one match in a
2929 line, it showed only the first of them. It now behaves in the same way as
2932 13. If the -o and -v options were combined for pcregrep, it printed a blank
2933 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2934 does the same. The return code can be used to tell if there were any
2937 14. Added --file-offsets and --line-offsets to pcregrep.
2939 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2940 infinitely looping recursion. The bug was that positive lookaheads were not
2941 being skipped when checking for a possible empty match (negative lookaheads
2942 and both kinds of lookbehind were skipped).
2944 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2945 inclusion of <windows.h> to before rather than after the definition of
2946 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2948 17. Specifying a possessive quantifier with a specific limit for a Unicode
2949 character property caused pcre_compile() to compile bad code, which led at
2950 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2951 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2952 caused the error; without that there was no problem.
2954 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2956 19. Added --enable-pcretest-libreadline.
2958 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2959 RE::GlobalReplace(). As a result, the number of replacements returned was
2960 double what it should be. I removed one of the increments, but Craig sent a
2961 later patch that removed the other one (the right fix) and added unit tests
2962 that check the return values (which was not done before).
2964 21. Several CMake things:
2966 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2967 the names libpcre and libpcreposix, not just pcre and pcreposix.
2969 (2) The above change means that pcretest and pcregrep are now correctly
2970 linked with the newly-built libraries, not previously installed ones.
2972 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2974 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2975 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2976 UTF-8 newline character). The key issue is that the pattern starts .*;
2977 this means that the match must be either at the beginning, or after a
2978 newline. The bug was in the code for advancing after a failed match and
2979 checking that the new position followed a newline. It was not taking
2980 account of UTF-8 characters correctly.
2982 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2983 character classes. PCRE was not treating the sequence [:...:] as a
2984 character class unless the ... were all letters. Perl, however, seems to
2985 allow any characters between [: and :], though of course it rejects as
2986 unknown any "names" that contain non-letters, because all the known class
2987 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2988 for example, whereas PCRE did not - it did not recognize a POSIX character
2989 class. This seemed a bit dangerous, so the code has been changed to be
2990 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2991 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2992 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2993 Perl does, and where it didn't before.
2995 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2996 Windows environments %n is disabled by default.
2999 Version 7.4 21-Sep-07
3000 ---------------------
3002 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
3003 means that a class such as [\s] counted as "explicit reference to CR or
3004 LF". That isn't really right - the whole point of the change was to try to
3005 help when there was an actual mention of one of the two characters. So now
3006 the change happens only if \r or \n (or a literal CR or LF) character is
3009 2. The 32-bit options word was also used for 6 internal flags, but the numbers
3010 of both had grown to the point where there were only 3 bits left.
3011 Fortunately, there was spare space in the data structure, and so I have
3012 moved the internal flags into a new 16-bit field to free up more option
3015 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3016 but did not set the internal JCHANGED flag - either of these is enough to
3017 control the way the "get" function works - but the PCRE_INFO_JCHANGED
3018 facility is supposed to tell if (?J) was ever used, so now (?J) at the
3019 start sets both bits.
3021 4. Added options (at build time, compile time, exec time) to change \R from
3022 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3024 5. doc/pcresyntax.html was missing from the distribution.
3026 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3027 compatibility, even though it is no longer used.
3029 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3030 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3031 windows.h file is present (where different names are used). [This was
3032 reversed later after testing - see 16 below.]
3034 8. Changed all #include <config.h> to #include "config.h". There were also
3035 some further <pcre.h> cases that I changed to "pcre.h".
3037 9. When pcregrep was used with the --colour option, it missed the line ending
3038 sequence off the lines that it output.
3040 10. It was pointed out to me that arrays of string pointers cause lots of
3041 relocations when a shared library is dynamically loaded. A technique of
3042 using a single long string with a table of offsets can drastically reduce
3043 these. I have refactored PCRE in four places to do this. The result is
3047 After changing UCP table: 187
3048 After changing error message table: 43
3049 After changing table of "verbs" 36
3050 After changing table of Posix names 22
3052 Thanks to the folks working on Gregex for glib for this insight.
3054 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3055 unicode-properties was also set.
3057 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3059 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3060 checked only for CRLF.
3062 14. Added casts to pcretest.c to avoid compiler warnings.
3064 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3066 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3067 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3068 entirely. This removes changes made in 7 above.
3070 17. The CMake files have been updated, and there is now more information about
3071 building with CMake in the NON-UNIX-USE document.
3074 Version 7.3 28-Aug-07
3075 ---------------------
3077 1. In the rejigging of the build system that eventually resulted in 7.1, the
3078 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3079 brackets there is not right, since it causes compilers to look for an
3080 installed pcre.h, not the version that is in the source that is being
3081 compiled (which of course may be different). I have changed it back to:
3085 I have a vague recollection that the change was concerned with compiling in
3086 different directories, but in the new build system, that is taken care of
3087 by the VPATH setting the Makefile.
3089 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3090 when the subject happened to end in the byte 0x85 (e.g. if the last
3091 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3092 characters but of course it shouldn't be taken as a newline when it is part
3093 of another character. The bug was that, for an unlimited repeat of . in
3094 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3095 characters when looking for a newline.
3097 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3099 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3102 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3103 long printing in the pcrecpp unittest when running under MinGW.
3105 6. ESC_K was left out of the EBCDIC table.
3107 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3108 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3109 limit also applies to "virtual nesting" when a pattern is recursive, and in
3110 this case 1000 isn't so big. I have been able to remove this limit at the
3111 expense of backing off one optimization in certain circumstances. Normally,
3112 when pcre_exec() would call its internal match() function recursively and
3113 immediately return the result unconditionally, it uses a "tail recursion"
3114 feature to save stack. However, when a subpattern that can match an empty
3115 string has an unlimited repetition quantifier, it no longer makes this
3116 optimization. That gives it a stack frame in which to save the data for
3117 checking that an empty string has been matched. Previously this was taken
3118 from the 1000-entry workspace that had been reserved. So now there is no
3119 explicit limit, but more stack is used.
3121 8. Applied Daniel's patches to solve problems with the import/export magic
3122 syntax that is required for Windows, and which was going wrong for the
3123 pcreposix and pcrecpp parts of the library. These were overlooked when this
3124 problem was solved for the main library.
3126 9. There were some crude static tests to avoid integer overflow when computing
3127 the size of patterns that contain repeated groups with explicit upper
3128 limits. As the maximum quantifier is 65535, the maximum group length was
3129 set at 30,000 so that the product of these two numbers did not overflow a
3130 32-bit integer. However, it turns out that people want to use groups that
3131 are longer than 30,000 bytes (though not repeat them that many times).
3132 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3133 made it possible to implement the integer overflow checks in a much more
3134 dynamic way, which I have now done. The artificial limitation on group
3135 length has been removed - we now have only the limit on the total length of
3136 the compiled pattern, which depends on the LINK_SIZE setting.
3138 10. Fixed a bug in the documentation for get/copy named substring when
3139 duplicate names are permitted. If none of the named substrings are set, the
3140 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3143 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3144 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3145 because the ] is interpreted as the first data character and the
3146 terminating ] is not found. PCRE has been made compatible with Perl in this
3147 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3148 cause memory overwriting.
3150 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3151 string has been matched (to stop an infinite loop). It was not recognizing
3152 a conditional subpattern that could match an empty string if that
3153 subpattern was within another subpattern. For example, it looped when
3154 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3155 condition was not nested. This bug has been fixed.
3157 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3158 past the start of the subject in the presence of bytes with the top bit
3159 set, for example "\x8aBCD".
3161 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3162 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3164 14. Optimized (?!) to (*FAIL).
3166 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3167 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3168 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3169 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3170 does: it's just the validity check that is more restrictive.
3172 16. Inserted checks for integer overflows during escape sequence (backslash)
3173 processing, and also fixed erroneous offset values for syntax errors during
3174 backslash processing.
3176 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3177 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3179 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3182 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3183 something other than just ASCII characters) inside a group that had an
3184 unlimited repeat caused a loop at compile time (while checking to see
3185 whether the group could match an empty string).
3187 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3188 [\P{Any}] did so. (Error in the code for printing property names.)
3190 21. An orphan \E inside a character class could cause a crash.
3192 22. A repeated capturing bracket such as (A)? could cause a wild memory
3193 reference during compilation.
3195 23. There are several functions in pcre_compile() that scan along a compiled
3196 expression for various reasons (e.g. to see if it's fixed length for look
3197 behind). There were bugs in these functions when a repeated \p or \P was
3198 present in the pattern. These operators have additional parameters compared
3199 with \d, etc, and these were not being taken into account when moving along
3200 the compiled data. Specifically:
3202 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3205 (b) An item such as \pL+ within a repeated group could cause crashes or
3208 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3209 "reference to non-existent subpattern" error.
3211 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3213 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3214 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3216 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3218 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3219 character were causing crashes (broken optimization).
3221 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3222 \p or \P) caused a compile-time loop.
3224 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3225 break. For example, the unstudied pattern [\r\n]A does not match the string
3226 "\r\nA" because change 7.0/46 below moves the current point on by two
3227 characters after failing to match at the start. However, the pattern \nA
3228 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3229 the same is true. There doesn't seem any very clean way out of this, but
3230 what I have chosen to do makes the common cases work: PCRE now takes note
3231 of whether there can be an explicit match for \r or \n anywhere in the
3232 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3233 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3234 pattern has explicit CR or LF references.
3236 29. Added (*CR) etc for changing newline setting at start of pattern.
3239 Version 7.2 19-Jun-07
3240 ---------------------
3242 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3243 which is apparently normally available under Windows.
3245 2. Re-jig the pcregrep tests with different newline settings in an attempt
3246 to make them independent of the local environment's newline setting.
3248 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3250 4. Some of the "internals" tests were previously cut out when the link size
3251 was not 2, because the output contained actual offsets. The recent new
3252 "Z" feature of pcretest means that these can be cut out, making the tests
3253 usable with all link sizes.
3255 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3256 stack recursion. This gives a massive performance boost under BSD, but just
3257 a small improvement under Linux. However, it saves one field in the frame
3260 6. Added more features from the forthcoming Perl 5.10:
3262 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3263 recursion call. It refers to the nth most recently opened parentheses.
3265 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3266 to be opened parentheses.
3268 (c) Conditions that refer to capturing parentheses can be specified
3269 relatively, for example, (?(-2)... or (?(+3)...
3271 (d) \K resets the start of the current match so that everything before
3274 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3276 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3279 (g) (?| introduces a group in which the numbering of parentheses in each
3280 alternative starts with the same number.
3282 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3284 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3287 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3288 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3289 for detecting groups that can match an empty string.
3291 9. A pattern with a very large number of alternatives (more than several
3292 hundred) was running out of internal workspace during the pre-compile
3293 phase, where pcre_compile() figures out how much memory will be needed. A
3294 bit of new cunning has reduced the workspace needed for groups with
3295 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3296 workspace instead of running out of the 4096 that are available.
3298 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3300 11. Applied patch from Google to remove an optimization that didn't quite work.
3301 The report of the bug said:
3303 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3304 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3305 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3307 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3308 it matched the wrong number of bytes.
3311 Version 7.1 24-Apr-07
3312 ---------------------
3314 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3315 that is more "standard", making use of automake and other Autotools. There
3316 is some re-arrangement of the files and adjustment of comments consequent
3319 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3320 for recursive directory scanning broke on some systems because the files
3321 are not scanned in any specific order and on different systems the order
3322 was different. A call to "sort" has been inserted into RunGrepTest for the
3323 approprate test as a short-term fix. In the longer term there may be an
3326 3. I had an email from Eric Raymond about problems translating some of PCRE's
3327 man pages to HTML (despite the fact that I distribute HTML pages, some
3328 people do their own conversions for various reasons). The problems
3329 concerned the use of low-level troff macros .br and .in. I have therefore
3330 removed all such uses from the man pages (some were redundant, some could
3331 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3332 HTML has been updated to handle .nf/.fi and to complain if it encounters
3335 4. Updated comments in configure.ac that get placed in config.h.in and also
3336 arranged for config.h to be included in the distribution, with the name
3337 config.h.generic, for the benefit of those who have to compile without
3338 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3340 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3341 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3342 makevp.bat for the current PCRE, using the additional files
3343 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3345 6. A Windows user reported a minor discrepancy with test 2, which turned out
3346 to be caused by a trailing space on an input line that had got lost in his
3347 copy. The trailing space was an accident, so I've just removed it.
3349 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3352 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3353 as "const" (a) because they are and (b) because it helps the PHP
3354 maintainers who have recently made a script to detect big data structures
3355 in the php code that should be moved to the .rodata section. I remembered
3356 to update Builducptable as well, so it won't revert if ucptable.h is ever
3359 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3360 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3361 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3362 support is not required. This saves 1.5-2K of code, which is important in
3365 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3366 so as not to refer to the tables, even though these functions will never be
3367 called when UTF-8 support is disabled. Otherwise there are problems with a
3370 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3372 (a) It was defining its arguments as char * instead of void *.
3374 (b) It was assuming that all moves were upwards in memory; this was true
3375 a long time ago when I wrote it, but is no longer the case.
3377 The emulated memove() is provided for those environments that have neither
3378 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3379 is clearly not the case, as these two bugs were recently reported.
3381 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3382 and Detrail to create the HTML documentation, the .txt form of the man
3383 pages, and it removes trailing spaces from listed files. It also creates
3384 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3385 case, it wraps all the #defines with #ifndefs. This script should be run
3388 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3389 with Unicode property support.
3391 (a) For a maximizing quantifier, if the two different cases of the
3392 character were of different lengths in their UTF-8 codings (there are
3393 some cases like this - I found 11), and the matching function had to
3394 back up over a mixture of the two cases, it incorrectly assumed they
3395 were both the same length.
3397 (b) When PCRE was configured to use the heap rather than the stack for
3398 recursion during matching, it was not correctly preserving the data for
3399 the other case of a UTF-8 character when checking ahead for a match
3400 while processing a minimizing repeat. If the check also involved
3401 matching a wide character, but failed, corruption could cause an
3402 erroneous result when trying to check for a repeat of the original
3405 13. Some tidying changes to the testing mechanism:
3407 (a) The RunTest script now detects the internal link size and whether there
3408 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3409 values substituted by "configure". (The RunGrepTest script already did
3410 this for UTF-8.) The configure.ac script no longer substitutes the
3413 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3414 with length and offset values. This means that the output is different
3415 for different internal link sizes. Test 2 is skipped for link sizes
3416 other than 2 because of this, bypassing the problem. Unfortunately,
3417 there was also a test in test 3 (the locale tests) that used /B and
3418 failed for link sizes other than 2. Rather than cut the whole test out,
3419 I have added a new /Z option to pcretest that replaces the length and
3420 offset values with spaces. This is now used to make test 3 independent
3421 of link size. (Test 2 will be tidied up later.)
3423 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3424 segmentation fault instead of returning the appropriate error message.
3426 15. In multiline mode when the newline sequence was set to "any", the pattern
3427 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3428 This doesn't seem right; it now treats the CRLF combination as the line
3429 ending, and so does not match in that case. It's only a pattern such as ^$
3430 that would hit this one: something like ^ABC$ would have failed after \r
3431 and then tried again after \r\n.
3433 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3434 in an attempt to make files that differ only in their line terminators
3435 compare equal. This works on Linux.
3437 17. Under certain error circumstances pcregrep might try to free random memory
3438 as it exited. This is now fixed, thanks to valgrind.
3440 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3441 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3442 was because its rules for how to advance for /g after matching an empty
3443 string at the end of a line did not allow for this case. They now check for
3446 20. pcretest is supposed to handle patterns and data of any length, by
3447 extending its buffers when necessary. It was getting this wrong when the
3448 buffer for a data line had to be extended.
3450 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3451 CRLF as a newline sequence.
3453 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3454 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3455 I have nevertheless tidied it up.
3457 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3459 24. Added a man page for pcre-config.
3462 Version 7.0 19-Dec-06
3463 ---------------------
3465 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3466 moving to gcc 4.1.1.
3468 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3469 sys/time.h, which is documented as needed for this function. It doesn't
3470 seem to matter on Linux, but it showed up on some releases of OS X.
3472 3. It seems that there are systems where bytes whose values are greater than
3473 127 match isprint() in the "C" locale. The "C" locale should be the
3474 default when a C program starts up. In most systems, only ASCII printing
3475 characters match isprint(). This difference caused the output from pcretest
3476 to vary, making some of the tests fail. I have changed pcretest so that:
3478 (a) When it is outputting text in the compiled version of a pattern, bytes
3479 other than 32-126 are always shown as hex escapes.
3481 (b) When it is outputting text that is a matched part of a subject string,
3482 it does the same, unless a different locale has been set for the match
3483 (using the /L modifier). In this case, it uses isprint() to decide.
3485 4. Fixed a major bug that caused incorrect computation of the amount of memory
3486 required for a compiled pattern when options that changed within the
3487 pattern affected the logic of the preliminary scan that determines the
3488 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3489 that the computed length was too small. The symptoms of this bug were
3490 either the PCRE error "internal error: code overflow" from pcre_compile(),
3491 or a glibc crash with a message such as "pcretest: free(): invalid next
3492 size (fast)". Examples of patterns that provoked this bug (shown in
3493 pcretest format) are:
3496 /(?x)(?-x: \s*#\s*)/
3500 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3501 is now done differently.
3503 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3504 wrapper classes; (b) implement a new function in the C++ scanner that is
3505 more efficient than the old way of doing things because it avoids levels of
3506 recursion in the regex matching; (c) add a paragraph to the documentation
3507 for the FullMatch() function.
3509 6. The escape sequence \n was being treated as whatever was defined as
3510 "newline". Not only was this contrary to the documentation, which states
3511 that \n is character 10 (hex 0A), but it also went horribly wrong when
3512 "newline" was defined as CRLF. This has been fixed.
3514 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3515 was being set to -1 for the "end of line" case (supposedly a value that no
3516 character can have). Though this value is never used (the check for end of
3517 line is "zero bytes in current character"), it caused compiler complaints.
3518 I've changed it to 0xffffffff.
3520 8. In pcre_version.c, the version string was being built by a sequence of
3521 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3522 string (as it is for production releases) called a macro with an empty
3523 argument. The C standard says the result of this is undefined. The gcc
3524 compiler treats it as an empty string (which was what was wanted) but it is
3525 reported that Visual C gives an error. The source has been hacked around to
3528 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3529 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3530 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3533 10. Originally, pcretest opened its input and output without "b"; then I was
3534 told that "b" was needed in some environments, so it was added for release
3535 5.0 to both the input and output. (It makes no difference on Unix-like
3536 systems.) Later I was told that it is wrong for the input on Windows. I've
3537 now abstracted the modes into two macros, to make it easier to fiddle with
3538 them, and removed "b" from the input mode under Windows.
3540 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3542 12. Added -help and --help to pcretest as an official way of being reminded
3545 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3546 and pcrecpp.cc because they annoy compilers at high warning levels.
3548 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3550 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3551 scripts are not C programs :-) and which was not noticed because it works
3554 16. pcretest is supposed to handle any length of pattern and data line (as one
3555 line or as a continued sequence of lines) by extending its input buffer if
3556 necessary. This feature was broken for very long pattern lines, leading to
3557 a string of junk being passed to pcre_compile() if the pattern was longer
3560 17. I have done a major re-factoring of the way pcre_compile() computes the
3561 amount of memory needed for a compiled pattern. Previously, there was code
3562 that made a preliminary scan of the pattern in order to do this. That was
3563 OK when PCRE was new, but as the facilities have expanded, it has become
3564 harder and harder to keep it in step with the real compile phase, and there
3565 have been a number of bugs (see for example, 4 above). I have now found a
3566 cunning way of running the real compile function in a "fake" mode that
3567 enables it to compute how much memory it would need, while actually only
3568 ever using a few hundred bytes of working memory and without too many
3569 tests of the mode. This should make future maintenance and development
3570 easier. A side effect of this work is that the limit of 200 on the nesting
3571 depth of parentheses has been removed (though this was never a serious
3572 limitation, I suspect). However, there is a downside: pcre_compile() now
3573 runs more slowly than before (30% or more, depending on the pattern). I
3574 hope this isn't a big issue. There is no effect on runtime performance.
3576 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3577 newline (only possible for the last line of a file) and it was a
3578 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3580 19. Added additional timing features to pcretest. (1) The -tm option now times
3581 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3582 separate command line item, by a number that specifies the number of
3583 repeats to use when timing. The default is 50000; this gives better
3584 precision, but takes uncomfortably long for very large patterns.
3586 20. Extended pcre_study() to be more clever in cases where a branch of a
3587 subpattern has no definite first character. For example, (a*|b*)[cd] would
3588 previously give no result from pcre_study(). Now it recognizes that the
3589 first character must be a, b, c, or d.
3591 21. There was an incorrect error "recursive call could loop indefinitely" if
3592 a subpattern (or the entire pattern) that was being tested for matching an
3593 empty string contained only one non-empty item after a nested subpattern.
3594 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3595 incorrectly, because the \d was being skipped in the check.
3597 22. The pcretest program now has a new pattern option /B and a command line
3598 option -b, which is equivalent to adding /B to every pattern. This causes
3599 it to show the compiled bytecode, without the additional information that
3600 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3601 is the same as /B/I).
3603 23. A new optimization is now able automatically to treat some sequences such
3604 as a*b as a*+b. More specifically, if something simple (such as a character
3605 or a simple class like \d) has an unlimited quantifier, and is followed by
3606 something that cannot possibly match the quantified thing, the quantifier
3607 is automatically "possessified".
3609 24. A recursive reference to a subpattern whose number was greater than 39
3610 went wrong under certain circumstances in UTF-8 mode. This bug could also
3611 have affected the operation of pcre_study().
3613 25. Realized that a little bit of performance could be had by replacing
3614 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3616 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3618 27. Possessive quantifiers such as a++ were previously implemented by turning
3619 them into atomic groups such as ($>a+). Now they have their own opcodes,
3620 which improves performance. This includes the automatically created ones
3623 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3624 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3625 the first matched character to be a colon. This applied both to named and
3628 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3630 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3631 building PCRE on Windows easier, so I have included it in the distribution.
3633 31. There is now a check in pcretest against a ridiculously large number being
3634 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3635 loop, the loop is abandoned.
3637 32. Forward references to subpatterns in conditions such as (?(2)...) where
3638 subpattern 2 is defined later cause pcre_compile() to search forwards in
3639 the pattern for the relevant set of parentheses. This search went wrong
3640 when there were unescaped parentheses in a character class, parentheses
3641 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3643 33. "Subroutine" calls and backreferences were previously restricted to
3644 referencing subpatterns earlier in the regex. This restriction has now
3647 34. Added a number of extra features that are going to be in Perl 5.10. On the
3648 whole, these are just syntactic alternatives for features that PCRE had
3649 previously implemented using the Python syntax or my own invention. The
3650 other formats are all retained for compatibility.
3652 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3653 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3654 also .NET compatible.
3656 (b) A recursion or subroutine call to a named group can now be defined as
3657 (?&name) as well as (?P>name).
3659 (c) A backreference to a named group can now be defined as \k<name> or
3660 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3661 5.10, are also .NET compatible.
3663 (d) A conditional reference to a named group can now use the syntax
3664 (?(<name>) or (?('name') as well as (?(name).
3666 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3667 groups (named and numbered) that are never evaluated inline, but can be
3668 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3669 is always false. There may be only one alternative in such a group.
3671 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3672 as the simple (?(R). The condition is true only if the most recent
3673 recursion is that of the given number or name. It does not search out
3674 through the entire recursion stack.
3676 (g) The escape \gN or \g{N} has been added, where N is a positive or
3677 negative number, specifying an absolute or relative reference.
3679 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3680 some "unreachable code" warnings.
3682 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3683 things, this adds five new scripts.
3685 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3686 There were also incompatibilities regarding the handling of \Q..\E inside
3687 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3688 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3690 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3691 matches an empty string, and forcibly breaks the loop. There were bugs in
3692 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3693 against aaaa the result was just "a" rather than "aaaa", for example. Two
3694 separate and independent bugs (that affected different cases) have been
3697 39. Refactored the code to abolish the use of different opcodes for small
3698 capturing bracket numbers. This is a tidy that I avoided doing when I
3699 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3700 The new approach is not only tidier, it makes it possible to reduce the
3701 memory needed to fix the previous bug (38).
3703 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3704 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3705 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3708 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3711 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3712 copy construction and assignment in the C++ wrapper.
3714 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3715 couple of bugs that could have given wrong results in the "--newline=crlf"
3718 44. Added a number of casts and did some reorganization of signed/unsigned int
3719 variables following suggestions from Dair Grant. Also renamed the variable
3720 "this" as "item" because it is a C++ keyword.
3722 45. Arranged for dftables to add
3724 #include "pcre_internal.h"
3726 to pcre_chartables.c because without it, gcc 4.x may remove the array
3727 definition from the final binary if PCRE is built into a static library and
3728 dead code stripping is activated.
3730 46. For an unanchored pattern, if a match attempt fails at the start of a
3731 newline sequence, and the newline setting is CRLF or ANY, and the next two
3732 characters are CRLF, advance by two characters instead of one.
3735 Version 6.7 04-Jul-06
3736 ---------------------
3738 1. In order to handle tests when input lines are enormously long, pcretest has
3739 been re-factored so that it automatically extends its buffers when
3740 necessary. The code is crude, but this _is_ just a test program. The
3741 default size has been increased from 32K to 50K.
3743 2. The code in pcre_study() was using the value of the re argument before
3744 testing it for NULL. (Of course, in any sensible call of the function, it
3747 3. The memmove() emulation function in pcre_internal.h, which is used on
3748 systems that lack both memmove() and bcopy() - that is, hardly ever -
3749 was missing a "static" storage class specifier.
3751 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3752 containing an extended class (one that cannot be represented by a bitmap
3753 because it contains high-valued characters or Unicode property items, e.g.
3754 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3755 pattern, but PCRE should not loop if you do not (it no longer does).
3756 [Detail: two cases were found: (a) a repeated subpattern containing an
3757 extended class; (b) a recursive reference to a subpattern that followed a
3758 previous extended class. It wasn't skipping over the extended class
3759 correctly when UTF-8 mode was not set.]
3761 5. A negated single-character class was not being recognized as fixed-length
3762 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3763 compile error "lookbehind assertion is not fixed length".
3765 6. The RunPerlTest auxiliary script was showing an unexpected difference
3766 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3767 write a Perl script that can interpret lines of an input file either as
3768 byte characters or as UTF-8, which is what "perltest" was being required to
3769 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3770 can't do is switch easily at run time between having the "use utf8;" pragma
3771 or not. In the end, I fudged it by using the RunPerlTest script to insert
3772 "use utf8;" explicitly for the UTF-8 tests.
3774 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3775 the end of the subject string, contrary to the documentation and to what
3776 Perl does. This was true of both matching functions. Now it matches only at
3777 the start of the subject and immediately after *internal* newlines.
3779 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3780 a pointer to an int instead of a pointer to an unsigned long int. This
3781 caused problems on 64-bit systems.
3783 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3784 instance of the 'standard' template library not being so standard".
3786 10. There was no check on the number of named subpatterns nor the maximum
3787 length of a subpattern name. The product of these values is used to compute
3788 the size of the memory block for a compiled pattern. By supplying a very
3789 long subpattern name and a large number of named subpatterns, the size
3790 computation could be caused to overflow. This is now prevented by limiting
3791 the length of names to 32 characters, and the number of named subpatterns
3794 11. Subpatterns that are repeated with specific counts have to be replicated in
3795 the compiled pattern. The size of memory for this was computed from the
3796 length of the subpattern and the repeat count. The latter is limited to
3797 65535, but there was no limit on the former, meaning that integer overflow
3798 could in principle occur. The compiled length of a repeated subpattern is
3799 now limited to 30,000 bytes in order to prevent this.
3801 12. Added the optional facility to have named substrings with the same name.
3803 13. Added the ability to use a named substring as a condition, using the
3804 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3805 are numbers (not recommended). Forward references are permitted.
3807 14. Added forward references in named backreferences (if you see what I mean).
3809 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3810 pattern could run off the end of the subject. For example, the pattern
3811 "(?s)(.{1,5})"8 did this with the subject "ab".
3813 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3814 PCRE_CASELESS was set when matching characters that were quantified with ?
3817 17. A character class other than a single negated character that had a minimum
3818 but no maximum quantifier - for example [ab]{6,} - was not handled
3819 correctly by pce_dfa_exec(). It would match only one character.
3821 18. A valid (though odd) pattern that looked like a POSIX character
3822 class but used an invalid character after [ (for example [[,abc,]]) caused
3823 pcre_compile() to give the error "Failed: internal error: code overflow" or
3824 in some cases to crash with a glibc free() error. This could even happen if
3825 the pattern terminated after [[ but there just happened to be a sequence of
3826 letters, a binary zero, and a closing ] in the memory that followed.
3828 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3829 over the years. Originally (before any Unicode support), just the bottom 8
3830 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3831 output from "man perlunicode" includes this:
3833 The regular expression compiler produces polymorphic opcodes. That
3834 is, the pattern adapts to the data and automatically switches to
3835 the Unicode character scheme when presented with Unicode data--or
3836 instead uses a traditional byte scheme when presented with byte
3839 Sadly, a wide octal escape does not cause a switch, and in a string with
3840 no other multibyte characters, these octal escapes are treated as before.
3841 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3842 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3845 I have not perpetrated such confusion in PCRE. Up till now, it took just
3846 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3847 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3848 translate to the appropriate multibyte character.
3850 29. Applied some refactoring to reduce the number of warnings from Microsoft
3851 and Borland compilers. This has included removing the fudge introduced
3852 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3853 a warning about an unused variable.
3855 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3856 characters since release 4.0, because Perl (from release 5.004) does not.
3857 [Or at least, is documented not to: some releases seem to be in conflict
3858 with the documentation.] However, when a pattern was studied with
3859 pcre_study() and all its branches started with \s, PCRE still included VT
3860 as a possible starting character. Of course, this did no harm; it just
3861 caused an unnecessary match attempt.
3863 22. Removed a now-redundant internal flag bit that recorded the fact that case
3864 dependency changed within the pattern. This was once needed for "required
3865 byte" processing, but is no longer used. This recovers a now-scarce options
3866 bit. Also moved the least significant internal flag bit to the most-
3867 significant bit of the word, which was not previously used (hangover from
3868 the days when it was an int rather than a uint) to free up another bit for
3871 23. Added support for CRLF line endings as well as CR and LF. As well as the
3872 default being selectable at build time, it can now be changed at runtime
3873 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3874 specify that it is scanning data with non-default line endings.
3876 24. Changed the definition of CXXLINK to make it agree with the definition of
3877 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3879 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3880 recursions. This makes a big different to stack usage for some patterns.
3882 26. If a subpattern containing a named recursion or subroutine reference such
3883 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3884 the space required for the compiled pattern went wrong and gave too small a
3885 value. Depending on the environment, this could lead to "Failed: internal
3886 error: code overflow at offset 49" or "glibc detected double free or
3889 27. Applied patches from Google (a) to support the new newline modes and (b) to
3890 advance over multibyte UTF-8 characters in GlobalReplace.
3892 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3893 difference for some implementation of PCRE in some Windows version.
3895 29. Added some extra testing facilities to pcretest:
3897 \q<number> in a data line sets the "match limit" value
3898 \Q<number> in a data line sets the "match recursion limt" value
3899 -S <number> sets the stack size, where <number> is in megabytes
3901 The -S option isn't available for Windows.
3904 Version 6.6 06-Feb-06
3905 ---------------------
3907 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3908 in pcreposix.h. I have copied the definition from pcre.h.
3910 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3911 because pcre.h is no longer a built file.
3913 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3914 not normally included in the compiled code.
3917 Version 6.5 01-Feb-06
3918 ---------------------
3920 1. When using the partial match feature with pcre_dfa_exec(), it was not
3921 anchoring the second and subsequent partial matches at the new starting
3922 point. This could lead to incorrect results. For example, with the pattern
3923 /1234/, partially matching against "123" and then "a4" gave a match.
3925 2. Changes to pcregrep:
3927 (a) All non-match returns from pcre_exec() were being treated as failures
3928 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3929 error message is output. Some extra information is given for the
3930 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3931 probably the only errors that are likely to be caused by users (by
3932 specifying a regex that has nested indefinite repeats, for instance).
3933 If there are more than 20 of these errors, pcregrep is abandoned.
3935 (b) A binary zero was treated as data while matching, but terminated the
3936 output line if it was written out. This has been fixed: binary zeroes
3937 are now no different to any other data bytes.
3939 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3940 used to set a locale for matching. The --locale=xxxx long option has
3941 been added (no short equivalent) to specify a locale explicitly on the
3942 pcregrep command, overriding the environment variables.
3944 (d) When -B was used with -n, some line numbers in the output were one less
3945 than they should have been.
3947 (e) Added the -o (--only-matching) option.
3949 (f) If -A or -C was used with -c (count only), some lines of context were
3950 accidentally printed for the final match.
3952 (g) Added the -H (--with-filename) option.
3954 (h) The combination of options -rh failed to suppress file names for files
3955 that were found from directory arguments.
3957 (i) Added the -D (--devices) and -d (--directories) options.
3959 (j) Added the -F (--fixed-strings) option.
3961 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3963 (l) Added the --colo(u)r option.
3965 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3966 is not present by default.
3968 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3969 items such as (?R) or (?1), when the recursion could match a number of
3970 alternatives. If it matched one of the alternatives, but subsequently,
3971 outside the recursion, there was a failure, the code tried to back up into
3972 the recursion. However, because of the way PCRE is implemented, this is not
3973 possible, and the result was an incorrect result from the match.
3975 In order to prevent this happening, the specification of recursion has
3976 been changed so that all such subpatterns are automatically treated as
3977 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3979 4. I had overlooked the fact that, in some locales, there are characters for
3980 which isalpha() is true but neither isupper() nor islower() are true. In
3981 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3982 and ordfeminine) are like this. This affected the treatment of \w and \W
3983 when they appeared in character classes, but not when they appeared outside
3984 a character class. The bit map for "word" characters is now created
3985 separately from the results of isalnum() instead of just taking it from the
3986 upper, lower, and digit maps. (Plus the underscore character, of course.)
3988 5. The above bug also affected the handling of POSIX character classes such as
3989 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3990 permanent tables. Instead, the bit maps for such a class were previously
3991 created as the appropriate unions of the upper, lower, and digit bitmaps.
3992 Now they are created by subtraction from the [[:word:]] class, which has
3995 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3996 It is created by subtracting the vertical space characters (\x09, \x0a,
3997 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3998 subtraction was done in the overall bitmap for a character class, meaning
3999 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
4000 be recognized. This bug has been fixed.
4002 7. Patches from the folks at Google:
4004 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
4005 real life, but is still worth protecting against".
4007 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
4008 regular expressions".
4010 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
4013 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
4014 "configure" and the latter not, in order to fix a problem somebody had
4015 with compiling the Arg class on HP-UX.
4017 (e) Improve the error-handling of the C++ wrapper a little bit.
4019 (f) New tests for checking recursion limiting.
4021 8. The pcre_memmove() function, which is used only if the environment does not
4022 have a standard memmove() function (and is therefore rarely compiled),
4023 contained two bugs: (a) use of int instead of size_t, and (b) it was not
4024 returning a result (though PCRE never actually uses the result).
4026 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
4027 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
4028 returned instead of calling malloc() with an overflowing number that would
4029 most likely cause subsequent chaos.
4031 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
4033 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
4034 with this option is matched, the nmatch and pmatch options of regexec() are
4037 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
4038 provided in case anyone wants to the the POSIX interface with UTF-8
4041 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
4042 C++ linking (needed for some HP-UX environments).
4044 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
4045 (unused parameter) and in the pcre_printint() function (omitted "default"
4046 switch label when the default is to do nothing).
4048 15. Added some code to make it possible, when PCRE is compiled as a C++
4049 library, to replace subject pointers for pcre_exec() with a smart pointer
4050 class, thus making it possible to process discontinuous strings.
4052 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
4053 much the same function. They were added by different people who were trying
4054 to make PCRE easy to compile on non-Unix systems. It has been suggested
4055 that PCRE_EXPORT be abolished now that there is more automatic apparatus
4056 for compiling on Windows systems. I have therefore replaced it with
4057 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
4058 defaults to "extern" for C or "extern C" for C++, which works fine on
4059 Unix-like systems. It is now possible to override the value of PCRE_DATA_
4060 SCOPE with something explicit in config.h. In addition:
4062 (a) pcreposix.h still had just "extern" instead of either of these macros;
4063 I have replaced it with PCRE_DATA_SCOPE.
4065 (b) Functions such as _pcre_xclass(), which are internal to the library,
4066 but external in the C sense, all had PCRE_EXPORT in their definitions.
4067 This is apparently wrong for the Windows case, so I have removed it.
4068 (It makes no difference on Unix-like systems.)
4070 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
4071 of recursive calls to match(). This is different to MATCH_LIMIT because
4072 that limits the total number of calls to match(), not all of which increase
4073 the depth of recursion. Limiting the recursion depth limits the amount of
4074 stack (or heap if NO_RECURSE is set) that is used. The default can be set
4075 when PCRE is compiled, and changed at run time. A patch from Google adds
4076 this functionality to the C++ interface.
4078 18. Changes to the handling of Unicode character properties:
4080 (a) Updated the table to Unicode 4.1.0.
4082 (b) Recognize characters that are not in the table as "Cn" (undefined).
4084 (c) I revised the way the table is implemented to a much improved format
4085 which includes recognition of ranges. It now supports the ranges that
4086 are defined in UnicodeData.txt, and it also amalgamates other
4087 characters into ranges. This has reduced the number of entries in the
4088 table from around 16,000 to around 3,000, thus reducing its size
4089 considerably. I realized I did not need to use a tree structure after
4090 all - a binary chop search is just as efficient. Having reduced the
4091 number of entries, I extended their size from 6 bytes to 8 bytes to
4092 allow for more data.
4094 (d) Added support for Unicode script names via properties such as \p{Han}.
4096 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4097 matching that character.
4099 20. When matching a repeated Unicode property with a minimum greater than zero,
4100 (for example \pL{2,}), PCRE could look past the end of the subject if it
4101 reached it while seeking the minimum number of characters. This could
4102 happen only if some of the characters were more than one byte long, because
4103 there is a check for at least the minimum number of bytes.
4105 21. Refactored the implementation of \p and \P so as to be more general, to
4106 allow for more different types of property in future. This has changed the
4107 compiled form incompatibly. Anybody with saved compiled patterns that use
4108 \p or \P will have to recompile them.
4110 22. Added "Any" and "L&" to the supported property types.
4112 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4113 but give a compile time error if the value is greater than 0xff.
4115 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4116 accidentally not being installed or uninstalled.
4118 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4119 made were to insert the current release number. This seemed silly, because
4120 it made things harder for people building PCRE on systems that don't run
4121 "configure". I have turned pcre.h into a distributed file, no longer built
4122 by "configure", with the version identification directly included. There is
4123 no longer a pcre.h.in file.
4125 However, this change necessitated a change to the pcre-config script as
4126 well. It is built from pcre-config.in, and one of the substitutions was the
4127 release number. I have updated configure.ac so that ./configure now finds
4128 the release number by grepping pcre.h.
4130 26. Added the ability to run the tests under valgrind.
4133 Version 6.4 05-Sep-05
4134 ---------------------
4136 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4137 "--" to be printed when multiple files were scanned, even when none of the
4138 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4139 consider it to be a bug, and have restored the previous behaviour.
4141 2. A couple of code tidies to get rid of compiler warnings.
4143 3. The pcretest program used to cheat by referring to symbols in the library
4144 whose names begin with _pcre_. These are internal symbols that are not
4145 really supposed to be visible externally, and in some environments it is
4146 possible to suppress them. The cheating is now confined to including
4147 certain files from the library's source, which is a bit cleaner.
4149 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4150 file's purpose clearer.
4152 5. Reorganized pcre_ucp_findchar().
4155 Version 6.3 15-Aug-05
4156 ---------------------
4158 1. The file libpcre.pc.in did not have general read permission in the tarball.
4160 2. There were some problems when building without C++ support:
4162 (a) If C++ support was not built, "make install" and "make test" still
4165 (b) There were problems when the value of CXX was explicitly set. Some
4166 changes have been made to try to fix these, and ...
4168 (c) --disable-cpp can now be used to explicitly disable C++ support.
4170 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4171 backslash in a target when C++ was disabled. This confuses some
4172 versions of "make", apparently. Using an intermediate variable solves
4173 this. (Same for CPP_LOBJ.)
4175 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4176 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4177 necessary on certain architectures.
4179 4. Added a setting of -export-symbols-regex to the link command to remove
4180 those symbols that are exported in the C sense, but actually are local
4181 within the library, and not documented. Their names all begin with
4182 "_pcre_". This is not a perfect job, because (a) we have to except some
4183 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4184 available (and never for static libraries). I have made a note to try to
4185 find a way round (a) in the future.
4188 Version 6.2 01-Aug-05
4189 ---------------------
4191 1. There was no test for integer overflow of quantifier values. A construction
4192 such as {1111111111111111} would give undefined results. What is worse, if
4193 a minimum quantifier for a parenthesized subpattern overflowed and became
4194 negative, the calculation of the memory size went wrong. This could have
4195 led to memory overwriting.
4197 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4199 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4200 operating environments where this matters.
4202 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4203 PCRE options from within the C++ wrapper.
4205 5. Named capturing subpatterns were not being correctly counted when a pattern
4206 was compiled. This caused two problems: (a) If there were more than 100
4207 such subpatterns, the calculation of the memory needed for the whole
4208 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4209 back references of the form \12, where the number was greater than 9, were
4210 not recognized as back references, even though there were sufficient
4211 previous subpatterns.
4213 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4214 versions of gcc, e.g. 2.95.4.
4217 Version 6.1 21-Jun-05
4218 ---------------------
4220 1. There was one reference to the variable "posix" in pcretest.c that was not
4221 surrounded by "#if !defined NOPOSIX".
4223 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4224 the cross-check on the old pcre_info() function, for the benefit of the
4225 cut-down version of PCRE that is currently imported into Exim.
4227 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4228 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4229 patterns that start (?i) but I don't think that matters. The use of (?i) is
4230 just an example; this all applies to the other options as well.
4232 4. Since libtool seems to echo the compile commands it is issuing, the output
4233 from "make" can be reduced a bit by putting "@" in front of each libtool
4236 5. Patch from the folks at Google for configure.in to be a bit more thorough
4237 in checking for a suitable C++ installation before trying to compile the
4238 C++ stuff. This should fix a reported problem when a compiler was present,
4239 but no suitable headers.
4241 6. The man pages all had just "PCRE" as their title. I have changed them to
4242 be the relevant file name. I have also arranged that these names are
4243 retained in the file doc/pcre.txt, which is a concatenation in text format
4244 of all the man pages except the little individual ones for each function.
4246 7. The NON-UNIX-USE file had not been updated for the different set of source
4247 files that come with release 6. I also added a few comments about the C++
4251 Version 6.0 07-Jun-05
4252 ---------------------
4254 1. Some minor internal re-organization to help with my DFA experiments.
4256 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4257 didn't matter for the library itself when fully configured, but did matter
4258 when compiling without UCP support, or within Exim, where the ucp files are
4261 3. Refactoring of the library code to split up the various functions into
4262 different source modules. The addition of the new DFA matching code (see
4263 below) to a single monolithic source would have made it really too
4264 unwieldy, quite apart from causing all the code to be include in a
4265 statically linked application, when only some functions are used. This is
4266 relevant even without the DFA addition now that patterns can be compiled in
4267 one application and matched in another.
4269 The downside of splitting up is that there have to be some external
4270 functions and data tables that are used internally in different modules of
4271 the library but which are not part of the API. These have all had their
4272 names changed to start with "_pcre_" so that they are unlikely to clash
4273 with other external names.
4275 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4276 a different (DFA) algorithm. Although it is slower than the original
4277 function, it does have some advantages for certain types of matching
4280 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4281 including restarting after a partial match.
4283 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4284 defined when compiling for Windows was sent to me. I have put it into the
4285 code, though I have no means of testing or verifying it.
4287 7. Added the pcre_refcount() auxiliary function.
4289 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4290 match before or at the first newline in the subject string. In pcretest,
4291 the /f option on a pattern can be used to set this.
4293 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4294 would behave wrongly. This has been present in PCRE since release 4.0.
4296 10. A number of changes to the pcregrep command:
4298 (a) Refactored how -x works; insert ^(...)$ instead of setting
4299 PCRE_ANCHORED and checking the length, in preparation for adding
4300 something similar for -w.
4302 (b) Added the -w (match as a word) option.
4304 (c) Refactored the way lines are read and buffered so as to have more
4305 than one at a time available.
4307 (d) Implemented a pcregrep test script.
4309 (e) Added the -M (multiline match) option. This allows patterns to match
4310 over several lines of the subject. The buffering ensures that at least
4311 8K, or the rest of the document (whichever is the shorter) is available
4312 for matching (and similarly the previous 8K for lookbehind assertions).
4314 (f) Changed the --help output so that it now says
4318 instead of two lines, one with "regex" and the other with "regexp"
4319 because that confused at least one person since the short forms are the
4320 same. (This required a bit of code, as the output is generated
4321 automatically from a table. It wasn't just a text change.)
4323 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4324 option but starts with a hyphen. Could be a pattern or a path name
4325 starting with a hyphen, for instance.
4327 (h) "-" can be given as a file name to represent stdin.
4329 (i) When file names are being printed, "(standard input)" is used for
4330 the standard input, for compatibility with GNU grep. Previously
4333 (j) The option --label=xxx can be used to supply a name to be used for
4334 stdin when file names are being printed. There is no short form.
4336 (k) Re-factored the options decoding logic because we are going to add
4337 two more options that take data. Such options can now be given in four
4338 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4340 (l) Added the -A, -B, and -C options for requesting that lines of context
4341 around matches be printed.
4343 (m) Added the -L option to print the names of files that do not contain
4344 any matching lines, that is, the complement of -l.
4346 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4347 continue to scan other files.
4349 (o) The -s option was incorrectly implemented. For compatibility with other
4350 greps, it now suppresses the error message for a non-existent or non-
4351 accessible file (but not the return code). There is a new option called
4352 -q that suppresses the output of matching lines, which was what -s was
4355 (p) Added --include and --exclude options to specify files for inclusion
4356 and exclusion when recursing.
4358 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4359 Hopefully, it now does.
4361 12. Missing cast in pcre_study().
4363 13. Added an "uninstall" target to the makefile.
4365 14. Replaced "extern" in the function prototypes in Makefile.in with
4366 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4367 world, but is set differently for Windows.
4369 15. Added a second compiling function called pcre_compile2(). The only
4370 difference is that it has an extra argument, which is a pointer to an
4371 integer error code. When there is a compile-time failure, this is set
4372 non-zero, in addition to the error test pointer being set to point to an
4373 error message. The new argument may be NULL if no error number is required
4374 (but then you may as well call pcre_compile(), which is now just a
4375 wrapper). This facility is provided because some applications need a
4376 numeric error indication, but it has also enabled me to tidy up the way
4377 compile-time errors are handled in the POSIX wrapper.
4379 16. Added VPATH=.libs to the makefile; this should help when building with one
4380 prefix path and installing with another. (Or so I'm told by someone who
4381 knows more about this stuff than I do.)
4383 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4384 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4385 match everything, including newlines. This is not POSIX-compatible, but
4386 somebody wanted the feature. From pcretest it can be activated by using
4387 both the P and the s flags.
4389 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4391 19. libpcre.pc was being incorrectly installed as executable.
4393 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4394 it now also looks for '\r' so that it will work unmodified on Windows.
4396 21. Added Google's contributed C++ wrapper to the distribution.
4398 22. Added some untidy missing memory free() calls in pcretest, to keep
4399 Electric Fence happy when testing.
4403 Version 5.0 13-Sep-04
4404 ---------------------
4406 1. Internal change: literal characters are no longer packed up into items
4407 containing multiple characters in a single byte-string. Each character
4408 is now matched using a separate opcode. However, there may be more than one
4409 byte in the character in UTF-8 mode.
4411 2. The pcre_callout_block structure has two new fields: pattern_position and
4412 next_item_length. These contain the offset in the pattern to the next match
4413 item, and its length, respectively.
4415 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4416 insertion of callouts before each pattern item. Added the /C option to
4417 pcretest to make use of this.
4419 4. On the advice of a Windows user, the lines
4421 #if defined(_WIN32) || defined(WIN32)
4422 _setmode( _fileno( stdout ), 0x8000 );
4423 #endif /* defined(_WIN32) || defined(WIN32) */
4425 have been added to the source of pcretest. This apparently does useful
4426 magic in relation to line terminators.
4428 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4429 for the benefit of those environments where the "b" makes a difference.
4431 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4432 to know about it. I have put a hack into configure.in that adds in code
4433 to set GCC=yes if CC=icc. This seems to end up at a point in the
4434 generated configure script that is early enough to affect the setting of
4435 compiler options, which is what is needed, but I have no means of testing
4436 whether it really works. (The user who reported this had patched the
4437 generated configure script, which of course I cannot do.)
4439 LATER: After change 22 below (new libtool files), the configure script
4440 seems to know about icc (and also ecc). Therefore, I have commented out
4441 this hack in configure.in.
4443 7. Added support for pkg-config (2 patches were sent in).
4445 8. Negated POSIX character classes that used a combination of internal tables
4446 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4447 [[:^ascii]]. Typically, they would match almost any characters. The other
4448 POSIX classes were not broken in this way.
4450 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4451 to find the match, as PCRE was deluded into thinking that the match had to
4452 start at the start point or following a newline. The same bug applied to
4453 patterns with negative forward assertions or any backward assertions
4454 preceding ".*" at the start, unless the pattern required a fixed first
4455 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4457 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4458 starting at the last subject character, bytes beyond the end of the subject
4461 11. Renamed the variable "class" as "classbits" to make life easier for C++
4462 users. (Previously there was a macro definition, but it apparently wasn't
4465 12. Added the new field "tables" to the extra data so that tables can be passed
4466 in at exec time, or the internal tables can be re-selected. This allows
4467 a compiled regex to be saved and re-used at a later time by a different
4468 program that might have everything at different addresses.
4470 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4471 -R library as well as a -L library.
4473 14. The debugging options of pcretest (-d on the command line or D on a
4474 pattern) showed incorrect output for anything following an extended class
4475 that contained multibyte characters and which was followed by a quantifier.
4477 15. Added optional support for general category Unicode character properties
4478 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4479 support. It adds about 90K to the size of the library. The meanings of the
4480 inbuilt class escapes such as \d and \s have NOT been changed.
4482 16. Updated pcredemo.c to include calls to free() to release the memory for the
4485 17. The generated file chartables.c was being created in the source directory
4486 instead of in the building directory. This caused the build to fail if the
4487 source directory was different from the building directory, and was
4490 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4491 file. No doubt somebody will tell me if they don't make sense... Also added
4492 Dan Mooney's comments about building on OpenVMS.
4494 19. Added support for partial matching via the PCRE_PARTIAL option for
4495 pcre_exec() and the \P data escape in pcretest.
4497 20. Extended pcretest with 3 new pattern features:
4499 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4500 write the compiled pattern to the file whose name is "rest-of-line".
4501 This is a straight binary dump of the data, with the saved pointer to
4502 the character tables forced to be NULL. The study data, if any, is
4503 written too. After writing, pcretest reads a new pattern.
4505 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4506 compiled pattern from the given file. There must not be any
4507 occurrences of "<" in the file name (pretty unlikely); if there are,
4508 pcretest will instead treat the initial "<" as a pattern delimiter.
4509 After reading in the pattern, pcretest goes on to read data lines as
4512 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4513 and 16-bit fields in a compiled pattern, to simulate a pattern that
4514 was compiled on a host of opposite endianness.
4516 21. The pcre-exec() function can now cope with patterns that were compiled on
4517 hosts of opposite endianness, with this restriction:
4519 As for any compiled expression that is saved and used later, the tables
4520 pointer field cannot be preserved; the extra_data field in the arguments
4521 to pcre_exec() should be used to pass in a tables address if a value
4522 other than the default internal tables were used at compile time.
4524 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4525 now diagnosed as an error. Previously, most of the time, a negative number
4526 would have been treated as zero, but if in addition "ovector" was passed as
4527 NULL, a crash could occur.
4529 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4530 new versions from the libtool 1.5 distribution (the last one is a copy of
4531 a file called libtool.m4). This seems to have fixed the need to patch
4532 "configure" to support Darwin 1.3 (which I used to do). However, I still
4533 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4536 24. Changed the PCRE licence to be the more standard "BSD" licence.
4539 Version 4.5 01-Dec-03
4540 ---------------------
4542 1. There has been some re-arrangement of the code for the match() function so
4543 that it can be compiled in a version that does not call itself recursively.
4544 Instead, it keeps those local variables that need separate instances for
4545 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4546 needs to "recurse". Keeping track of where control must go is done by means
4547 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4548 hide most of the details from the main code, and operates only if
4549 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4550 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4553 To make it easier for callers to provide specially tailored get/free
4554 functions for this usage, two new functions, pcre_stack_malloc, and
4555 pcre_stack_free, are used. They are always called in strict stacking order,
4556 and the size of block requested is always the same.
4558 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4559 PCRE has been compiled to use the stack or the heap for recursion. The
4560 -C option of pcretest uses this to show which version is compiled.
4562 A new data escape \S, is added to pcretest; it causes the amounts of store
4563 obtained and freed by both kinds of malloc/free at match time to be added
4566 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4567 what's available on my current Linux desktop machine.
4569 3. When matching a UTF-8 string, the test for a valid string at the start has
4570 been extended. If start_offset is not zero, PCRE now checks that it points
4571 to a byte that is the start of a UTF-8 character. If not, it returns
4572 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4573 this is necessary because there may be backward assertions in the pattern.
4574 When matching the same subject several times, it may save resources to use
4575 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4577 4. The code for checking the validity of UTF-8 strings has been tightened so
4578 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4579 containing "overlong sequences".
4581 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4582 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4583 should have been "&", but it just so happened that all the cases this let
4584 through by mistake were picked up later in the function.
4586 6. I had used a variable called "isblank" - this is a C99 function, causing
4587 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4589 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4590 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4591 so the version is shown; (c) stop "make test" echoing "./RunTest".
4593 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4595 9. The prototype for memmove() for systems that don't have it was using
4596 size_t, but the inclusion of the header that defines size_t was later. I've
4597 moved the #includes for the C headers earlier to avoid this.
4599 10. Added some adjustments to the code to make it easier to compiler on certain
4602 (a) Some "const" qualifiers were missing.
4603 (b) Added the macro EXPORT before all exported functions; by default this
4604 is defined to be empty.
4605 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4606 that it reads its output file name as an argument instead of writing
4607 to the standard output and assuming this can be redirected.
4609 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4610 class containing characters with values greater than 255, PCRE compilation
4613 12. A recursive reference to a subpattern that was within another subpattern
4614 that had a minimum quantifier of zero caused PCRE to crash. For example,
4615 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4616 recursion. If the recursively-called subpattern itself had a zero repeat,
4619 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4620 buffer into which it was copied (for escape processing) was still set at
4621 1024, so long lines caused crashes.
4623 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4624 "internal error: code overflow...". This applied to any character class
4625 that was followed by a possessive quantifier.
4627 15. Modified the Makefile to add libpcre.la as a prerequisite for
4628 libpcreposix.la because I was told this is needed for a parallel build to
4631 16. If a pattern that contained .* following optional items at the start was
4632 studied, the wrong optimizing data was generated, leading to matching
4633 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4634 matching string must start with a or b or c. The correct conclusion for
4635 this pattern is that a match can start with any character.
4638 Version 4.4 13-Aug-03
4639 ---------------------
4641 1. In UTF-8 mode, a character class containing characters with values between
4642 127 and 255 was not handled correctly if the compiled pattern was studied.
4643 In fixing this, I have also improved the studying algorithm for such
4646 2. Three internal functions had redundant arguments passed to them. Removal
4647 might give a very teeny performance improvement.
4649 3. Documentation bug: the value of the capture_top field in a callout is *one
4650 more than* the number of the hightest numbered captured substring.
4652 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4653 in incorrectly linking with a previously installed version. They now link
4654 explicitly with libpcre.la.
4656 5. configure.in no longer needs to recognize Cygwin specially.
4658 6. A problem in pcre.in for Windows platforms is fixed.
4660 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4661 pcretest, it used to include the size of the study block as part of its
4662 output. Unfortunately, the structure contains a field that has a different
4663 size on different hardware architectures. This meant that the tests that
4664 showed this size failed. As the block is currently always of a fixed size,
4665 this information isn't actually particularly useful in pcretest output, so
4666 I have just removed it.
4668 8. Three pre-processor statements accidentally did not start in column 1.
4669 Sadly, there are *still* compilers around that complain, even though
4670 standard C has not required this for well over a decade. Sigh.
4672 9. In pcretest, the code for checking callouts passed small integers in the
4673 callout_data field, which is a void * field. However, some picky compilers
4674 complained about the casts involved for this on 64-bit systems. Now
4675 pcretest passes the address of the small integer instead, which should get
4676 rid of the warnings.
4678 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4679 both compile and run time, and gives an error if an invalid UTF-8 sequence
4680 is found. There is a option for disabling this check in cases where the
4681 string is known to be correct and/or the maximum performance is wanted.
4683 11. In response to a bug report, I changed one line in Makefile.in from
4685 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4687 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4689 to look similar to other lines, but I have no way of telling whether this
4690 is the right thing to do, as I do not use Windows. No doubt I'll get told
4694 Version 4.3 21-May-03
4695 ---------------------
4697 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4700 2. Some refactoring to improve the quality of the code:
4702 (i) The utf8_table... variables are now declared "const".
4704 (ii) The code for \cx, which used the "case flipping" table to upper case
4705 lower case letters, now just substracts 32. This is ASCII-specific,
4706 but the whole concept of \cx is ASCII-specific, so it seems
4709 (iii) PCRE was using its character types table to recognize decimal and
4710 hexadecimal digits in the pattern. This is silly, because it handles
4711 only 0-9, a-f, and A-F, but the character types table is locale-
4712 specific, which means strange things might happen. A private
4713 table is now used for this - though it costs 256 bytes, a table is
4714 much faster than multiple explicit tests. Of course, the standard
4715 character types table is still used for matching digits in subject
4718 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4719 ending in _t are). So I've renamed it as ESC_tee.
4721 3. The first argument for regexec() in the POSIX wrapper should have been
4724 4. Changed pcretest to use malloc() for its buffers so that they can be
4725 Electric Fenced for debugging.
4727 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4728 to read one or more bytes before the start of the subject string. Often this
4729 had no effect on PCRE's behaviour, but in some circumstances it could
4730 provoke a segmentation fault.
4732 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4733 to try to read one or more bytes before the start of the subject string.
4735 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4736 UTF-8 support could misbehave in various ways if the subject string
4737 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4738 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4739 back over UTF-8 characters.)
4742 Version 4.2 14-Apr-03
4743 ---------------------
4745 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4747 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4748 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4749 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4750 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4751 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4753 Note: automatic setting of the BUILD variables is not yet working
4754 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4755 compile-time but not at link-time
4756 [LINK]: use for linking executables only
4757 make different versions for Windows and non-Windows
4758 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4760 [LINK_FOR_BUILD]: new variable
4761 [OBJEXT]: use throughout
4762 [EXEEXT]: use throughout
4763 <winshared>: new target
4764 <wininstall>: new target
4765 <dftables.o>: use native compiler
4766 <dftables>: use native linker
4767 <install>: handle Windows platform correctly
4770 copy DLL to top builddir before testing
4772 As part of these changes, -no-undefined was removed again. This was reported
4773 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4776 3. Some tidies to get rid of compiler warnings:
4778 . In the match_data structure, match_limit was an unsigned long int, whereas
4779 match_call_count was an int. I've made them both unsigned long ints.
4781 . In pcretest the fact that a const uschar * doesn't automatically cast to
4782 a void * provoked a warning.
4784 . Turning on some more compiler warnings threw up some "shadow" variables
4785 and a few more missing casts.
4787 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4788 option, a class that contained a single character with a value between 128
4789 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4791 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4792 option, a class that contained several characters, but with at least one
4793 whose value was between 128 and 255 caused PCRE to crash.
4796 Version 4.1 12-Mar-03
4797 ---------------------
4799 1. Compiling with gcc -pedantic found a couple of places where casts were
4800 needed, and a string in dftables.c that was longer than standard compilers are
4801 required to support.
4803 2. Compiling with Sun's compiler found a few more places where the code could
4804 be tidied up in order to avoid warnings.
4806 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4807 first of these names is deprecated in the latest Autoconf in favour of the name
4808 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4809 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4810 analogy I have changed it to CFLAGS_FOR_BUILD.
4812 4. Added -no-undefined to the linking command in the Makefile, because this is
4813 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4814 linking step for the pcreposix library.
4816 5. PCRE was failing to diagnose the case of two named groups with the same
4819 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4820 literal character that is needed in the subject for a match, and scans along to
4821 ensure that it is present before embarking on the full matching process. This
4822 saves time in cases of nested unlimited repeats that are never going to match.
4823 Problem: the scan can take a lot of time if the subject is very long (e.g.
4824 megabytes), thus penalizing straightforward matches. It is now done only if the
4825 amount of subject to be scanned is less than 1000 bytes.
4827 7. A lesser problem with the same optimization is that it was recording the
4828 first character of an anchored pattern as "needed", thus provoking a search
4829 right along the subject, even when the first match of the pattern was going to
4830 fail. The "needed" character is now not set for anchored patterns, unless it
4831 follows something in the pattern that is of non-fixed length. Thus, it still
4832 fulfils its original purpose of finding quick non-matches in cases of nested
4833 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4836 Version 4.0 17-Feb-03
4837 ---------------------
4839 1. If a comment in an extended regex that started immediately after a meta-item
4840 extended to the end of string, PCRE compiled incorrect data. This could lead to
4841 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4843 2. Moved to autoconf 2.53 and libtool 1.4.2.
4845 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4846 the special perltest8 script is no longer needed - all the tests can be run
4847 from a single perltest script.
4849 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4850 by \s. It has now been removed in PCRE. This means it isn't recognized as
4851 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4852 class [:space:] *does* include VT, thereby creating a mess.
4854 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4857 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4858 its new features into the main test script, reducing the number of scripts.
4860 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4861 were backward compatible, and made the (?i) apply to the whole pattern, as if
4862 /i were given. Now it behaves more logically, and applies the option setting
4863 only to what follows. PCRE has been changed to follow suit. However, if it
4864 finds options settings right at the start of the pattern, it extracts them into
4865 the global options, as before. Thus, they show up in the info data.
4867 8. Added support for the \Q...\E escape sequence. Characters in between are
4868 treated as literals. This is slightly different from Perl in that $ and @ are
4869 also handled as literals inside the quotes. In Perl, they will cause variable
4870 interpolation. Note the following examples:
4872 Pattern PCRE matches Perl matches
4874 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4875 \Qabc\$xyz\E abc\$xyz abc\$xyz
4876 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4878 For compatibility with Perl, \Q...\E sequences are recognized inside character
4879 classes as well as outside them.
4881 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4882 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4883 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4884 signed/unsigned warnings.
4886 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4887 option for pcretest, so I've replaced it by a simple function that does just
4890 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4893 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4894 Java package. This provides some syntactic sugar for simple cases of what my
4895 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4896 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4897 item, you can use this simplified notation. Note that only makes sense with
4898 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4899 greediness, whatever the setting of the PCRE_UNGREEDY option.
4901 13. A change of greediness default within a pattern was not taking effect at
4902 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4903 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4904 was abstracted outside.
4906 14. PCRE now supports the \G assertion. It is true when the current matching
4907 position is at the start point of the match. This differs from \A when the
4908 starting offset is non-zero. Used with the /g option of pcretest (or similar
4909 code), it works in the same way as it does for Perl's /g option. If all
4910 alternatives of a regex begin with \G, the expression is anchored to the start
4911 match position, and the "anchored" flag is set in the compiled expression.
4913 15. Some bugs concerning the handling of certain option changes within patterns
4914 have been fixed. These applied to options other than (?ims). For example,
4915 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4916 been the other way round. Some of this was related to change 7 above.
4918 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4919 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4920 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4921 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4923 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4924 mode. Unlike ".", it always matches newline, whatever the setting of
4925 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4926 assertions. Perl allows it, but it doesn't (in general) work because it can't
4927 calculate the length of the lookbehind. At least, that's the case for Perl
4928 5.8.0 - I've been told they are going to document that it doesn't work in
4931 18. Added an error diagnosis for escapes that PCRE does not support: these are
4932 \L, \l, \N, \P, \p, \U, \u, and \X.
4934 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4935 reading past the end of the pattern in cases such as /[abcd/.
4937 20. PCRE was getting more memory than necessary for patterns with classes that
4938 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4940 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4941 compiling PCRE for use with Virtual Pascal.
4943 22. Small fix to the Makefile to make it work properly if the build is done
4944 outside the source tree.
4946 23. Added a new extension: a condition to go with recursion. If a conditional
4947 subpattern starts with (?(R) the "true" branch is used if recursion has
4948 happened, whereas the "false" branch is used only at the top level.
4950 24. When there was a very long string of literal characters (over 255 bytes
4951 without UTF support, over 250 bytes with UTF support), the computation of how
4952 much memory was required could be incorrect, leading to segfaults or other
4955 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4956 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4957 there was a subsequent back reference to those brackets. This meant that, for
4958 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4959 possible to check for precisely this case. All we can do is abandon the
4960 optimization if .* occurs inside capturing brackets when there are any back
4961 references whatsoever. (See below for a better fix that came later.)
4963 26. The handling of the optimization for finding the first character of a
4964 non-anchored pattern, and for finding a character that is required later in the
4965 match were failing in some cases. This didn't break the matching; it just
4966 failed to optimize when it could. The way this is done has been re-implemented.
4968 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4970 28. Added a new feature that provides some of the functionality that Perl
4971 provides with (?{...}). The facility is termed a "callout". The way it is done
4972 in PCRE is for the caller to provide an optional function, by setting
4973 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4974 global variable. By default it is unset, which disables all calling out. To get
4975 the function called, the regex must include (?C) at appropriate points. This
4976 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4977 This provides a means of identifying different callout points. When PCRE
4978 reaches such a point in the regex, if pcre_callout has been set, the external
4979 function is called. It is provided with data in a structure called
4980 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4981 matching continues; if it returns a non-zero value, the match at the current
4982 point fails. However, backtracking will occur if possible. [This was changed
4983 later and other features added - see item 49 below.]
4985 29. pcretest is upgraded to test the callout functionality. It provides a
4986 callout function that displays information. By default, it shows the start of
4987 the match and the current position in the text. There are some new data escapes
4988 to vary what happens:
4990 \C+ in addition, show current contents of captured substrings
4991 \C- do not supply a callout function
4992 \C!n return 1 when callout number n is reached
4993 \C!n!m return 1 when callout number n is reached for the mth time
4995 30. If pcregrep was called with the -l option and just a single file name, it
4996 output "<stdin>" if a match was found, instead of the file name.
4998 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4999 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
5000 pcre_exec(). This saves a malloc/free per call. The default value of
5001 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
5004 32. The default maximum size of a compiled pattern is 64K. There have been a
5005 few cases of people hitting this limit. The code now uses macros to handle the
5006 storing of links as offsets within the compiled pattern. It defaults to 2-byte
5007 links, but this can be changed to 3 or 4 bytes by --with-link-size when
5008 configuring. Tests 2 and 5 work only with 2-byte links because they output
5009 debugging information about compiled patterns.
5011 33. Internal code re-arrangements:
5013 (a) Moved the debugging function for printing out a compiled regex into
5014 its own source file (printint.c) and used #include to pull it into
5015 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
5018 (b) Defined the list of op-code names for debugging as a macro in
5019 internal.h so that it is next to the definition of the opcodes.
5021 (c) Defined a table of op-code lengths for simpler skipping along compiled
5022 code. This is again a macro in internal.h so that it is next to the
5023 definition of the opcodes.
5025 34. Added support for recursive calls to individual subpatterns, along the
5026 lines of Robin Houston's patch (but implemented somewhat differently).
5028 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
5029 allow it to read and process whole directories in Win32. This code was
5030 contributed by Lionel Fourquaux; it has not been tested by me.
5032 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
5033 used to name a group. Names consist of alphanumerics and underscores, and must
5034 be unique. Back references use the syntax (?P=name) and recursive calls use
5035 (?P>name) which is a PCRE extension to the Python extension. Groups still have
5036 numbers. The function pcre_fullinfo() can be used after compilation to extract
5037 a name/number map. There are three relevant calls:
5039 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
5040 PCRE_INFO_NAMECOUNT yields the number of entries
5041 PCRE_INFO_NAMETABLE yields a pointer to the map.
5043 The map is a vector of fixed-size entries. The size of each entry depends on
5044 the length of the longest name used. The first two bytes of each entry are the
5045 group number, most significant byte first. There follows the corresponding
5046 name, zero terminated. The names are in alphabetical order.
5048 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
5049 case instead of 255. Making it the same both with and without UTF-8 support
5050 means that the same test output works with both.
5052 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
5053 calling malloc() with a zero argument.
5055 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
5056 optimization. I've improved things by keeping a bitmap of backreferences with
5057 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
5058 fact referenced, the optimization can be applied. It is unlikely that a
5059 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
5060 the match to follow \n) will appear inside brackets with a number greater than
5061 31, but if it does, any back reference > 31 suppresses the optimization.
5063 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
5064 of disabling numbered capturing parentheses. Any opening parenthesis that is
5065 not followed by ? behaves as if it were followed by ?: but named parentheses
5066 can still be used for capturing (and they will acquire numbers in the usual
5069 41. Redesigned the return codes from the match() function into yes/no/error so
5070 that errors can be passed back from deep inside the nested calls. A malloc
5071 failure while inside a recursive subpattern call now causes the
5072 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
5074 42. It is now possible to set a limit on the number of times the match()
5075 function is called in a call to pcre_exec(). This facility makes it possible to
5076 limit the amount of recursion and backtracking, though not in a directly
5077 obvious way, because the match() function is used in a number of different
5078 circumstances. The count starts from zero for each position in the subject
5079 string (for non-anchored patterns). The default limit is, for compatibility, a
5080 large number, namely 10 000 000. You can change this in two ways:
5082 (a) When configuring PCRE before making, you can use --with-match-limit=n
5083 to set a default value for the compiled library.
5085 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5086 a different value is set. See 45 below.
5088 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5090 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5091 of things that can be changed at compile time. The first argument specifies
5092 what is wanted and the second points to where the information is to be placed.
5093 The current list of available information is:
5097 The output is an integer that is set to one if UTF-8 support is available;
5098 otherwise it is set to zero.
5102 The output is an integer that it set to the value of the code that is used for
5103 newline. It is either LF (10) or CR (13).
5105 PCRE_CONFIG_LINK_SIZE
5107 The output is an integer that contains the number of bytes used for internal
5108 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5110 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
5112 The output is an integer that contains the threshold above which the POSIX
5113 interface uses malloc() for output vectors. See item 31 above.
5115 PCRE_CONFIG_MATCH_LIMIT
5117 The output is an unsigned integer that contains the default limit of the number
5118 of match() calls in a pcre_exec() execution. See 42 above.
5120 44. pcretest has been upgraded by the addition of the -C option. This causes it
5121 to extract all the available output from the new pcre_config() function, and to
5122 output it. The program then exits immediately.
5124 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5125 order to support additional features. One way would have been to define
5126 pcre_exec2() (for example) with extra arguments, but this would not have been
5127 extensible, and would also have required all calls to the original function to
5128 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5129 is used for passing in "extra" data from pcre_study().
5131 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5132 contains the following fields:
5134 flags a bitmap indicating which of the following fields are set
5135 study_data opaque data from pcre_study()
5136 match_limit a way of specifying a limit on match() calls for a specific
5138 callout_data data for callouts (see 49 below)
5140 The flag bits are also defined in pcre.h, and are
5142 PCRE_EXTRA_STUDY_DATA
5143 PCRE_EXTRA_MATCH_LIMIT
5144 PCRE_EXTRA_CALLOUT_DATA
5146 The pcre_study() function now returns one of these new pcre_extra blocks, with
5147 the actual study data pointed to by the study_data field, and the
5148 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5149 before. That is, this change is entirely upwards-compatible and requires no
5150 change to existing code.
5152 If you want to pass in additional data to pcre_exec(), you can either place it
5153 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5156 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5157 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5158 times with different match limits, until it finds the minimum value needed for
5159 pcre_exec() to complete. The value is then output. This can be instructive; for
5160 most simple matches the number is quite small, but for pathological cases it
5161 gets very large very quickly.
5163 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5164 returns the size of the data block pointed to by the study_data field in a
5165 pcre_extra block, that is, the value that was passed as the argument to
5166 pcre_malloc() when PCRE was getting memory in which to place the information
5167 created by pcre_study(). The fourth argument should point to a size_t variable.
5168 pcretest has been extended so that this information is shown after a successful
5169 pcre_study() call when information about the compiled regex is being displayed.
5171 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5172 because what follows is always an absolute path. (Later: it turns out that this
5173 is more than cosmetic for MinGW, because it doesn't like empty path
5176 49. Some changes have been made to the callout feature (see 28 above):
5178 (i) A callout function now has three choices for what it returns:
5180 0 => success, carry on matching
5181 > 0 => failure at this point, but backtrack if possible
5182 < 0 => serious error, return this value from pcre_exec()
5184 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5185 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5186 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5187 use by callout functions. It will never be used by PCRE itself.
5189 (ii) The pcre_extra structure (see 45 above) has a void * field called
5190 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5191 pcre_callout_block structure has a field of the same name. The contents of
5192 the field passed in the pcre_extra structure are passed to the callout
5193 function in the corresponding field in the callout block. This makes it
5194 easier to use the same callout-containing regex from multiple threads. For
5195 testing, the pcretest program has a new data escape
5197 \C*n pass the number n (may be negative) as callout_data
5199 If the callout function in pcretest receives a non-zero value as
5200 callout_data, it returns that value.
5202 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5203 there were some redundant $(CFLAGS) in commands that are now specified as
5204 $(LINK), which already includes $(CFLAGS).
5206 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5207 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5208 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5209 one-byte characters throughout. Note that case-insensitive matching applies
5210 only to characters whose values are less than 256. PCRE doesn't support the
5211 notion of cases for higher-valued characters.
5213 (i) A character class whose characters are all within 0-255 is handled as
5214 a bit map, and the map is inverted for negative classes. Previously, a
5215 character > 255 always failed to match such a class; however it should
5216 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5218 (ii) A negated character class with a single character < 255 is coded as
5219 "not this character" (OP_NOT). This wasn't working properly when the test
5220 character was multibyte, either singly or repeated.
5222 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5223 mode, for example: \x{100}{2,3}.
5225 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5226 singly or repeated) now correctly test multibyte characters. However,
5227 PCRE doesn't recognize any characters with values greater than 255 as
5228 digits, spaces, or word characters. Such characters always match \D, \S,
5229 and \W, and never match \d, \s, or \w.
5231 (v) Classes may now contain characters and character ranges with values
5232 greater than 255. For example: [ab\x{100}-\x{400}].
5234 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5237 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5238 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5239 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5242 53. The single man page has become too large. I have therefore split it up into
5243 a number of separate man pages. These also give rise to individual HTML pages;
5244 these are now put in a separate directory, and there is an index.html page that
5245 lists them all. Some hyperlinking between the pages has been installed.
5247 54. Added convenience functions for handling named capturing parentheses.
5249 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5250 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5251 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5254 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5255 calling configure. These values are used when compiling the dftables.c program
5256 which is run to generate the source of the default character tables. They
5257 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5258 you will need to set these values.
5260 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5263 Version 3.9 02-Jan-02
5264 ---------------------
5266 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5268 2. If --disable-static was given, the building process failed when trying to
5269 build pcretest and pcregrep. (For some reason it was using libtool to compile
5270 them, which is not right, as they aren't part of the library.)
5273 Version 3.8 18-Dec-01
5274 ---------------------
5276 1. The experimental UTF-8 code was completely screwed up. It was packing the
5277 bytes in the wrong order. How dumb can you get?
5280 Version 3.7 29-Oct-01
5281 ---------------------
5283 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5284 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5285 this didn't happen under Solaris 8, where I normally test things.
5287 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5288 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5289 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5290 this fix an hour or so after the initial 3.7 release.)
5293 Version 3.6 23-Oct-01
5294 ---------------------
5296 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5297 offsets passed as NULL with zero offset count.
5299 2. The config.guess and config.sub files had not been updated when I moved to
5300 the latest autoconf.
5303 Version 3.5 15-Aug-01
5304 ---------------------
5306 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5309 2. By using declared but undefined structures, we can avoid using "void"
5310 definitions in pcre.h while keeping the internal definitions of the structures
5313 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5314 user point of view, this means that both static and shared libraries are built
5315 by default, but this can be individually controlled. More of the work of
5316 handling this static/shared cases is now inside libtool instead of PCRE's make
5319 4. The pcretest utility is now installed along with pcregrep because it is
5320 useful for users (to test regexs) and by doing this, it automatically gets
5321 relinked by libtool. The documentation has been turned into a man page, so
5322 there are now .1, .txt, and .html versions in /doc.
5324 5. Upgrades to pcregrep:
5325 (i) Added long-form option names like gnu grep.
5326 (ii) Added --help to list all options with an explanatory phrase.
5327 (iii) Added -r, --recursive to recurse into sub-directories.
5328 (iv) Added -f, --file to read patterns from a file.
5330 6. pcre_exec() was referring to its "code" argument before testing that
5331 argument for NULL (and giving an error if it was NULL).
5333 7. Upgraded Makefile.in to allow for compiling in a different directory from
5334 the source directory.
5336 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5337 options bits, the pointer it was passed was to an int instead of to an unsigned
5338 long int. This mattered only on 64-bit systems.
5340 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5341 generated) instead of pcre.in, which it its source. Also made the same change
5342 in several of the .c files.
5344 10. A new release of gcc defines printf() as a macro, which broke pcretest
5345 because it had an ifdef in the middle of a string argument for printf(). Fixed
5346 by using separate calls to printf().
5348 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5349 script, to force use of CR or LF instead of \n in the source. On non-Unix
5350 systems, the value can be set in config.h.
5352 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5353 absolute limit. Changed the text of the error message to make this clear, and
5354 likewise updated the man page.
5356 13. The limit of 99 on the number of capturing subpatterns has been removed.
5357 The new limit is 65535, which I hope will not be a "real" limit.
5360 Version 3.4 22-Aug-00
5361 ---------------------
5363 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5365 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5368 Version 3.3 01-Aug-00
5369 ---------------------
5371 1. If an octal character was given, but the value was greater than \377, it
5372 was not getting masked to the least significant bits, as documented. This could
5373 lead to crashes in some systems.
5375 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5376 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5378 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5379 These just pass their arguments on to (pcre_free)(), but they are provided
5380 because some uses of PCRE bind it to non-C systems that can call its functions,
5381 but cannot call free() or pcre_free() directly.
5383 4. Add "make test" as a synonym for "make check". Corrected some comments in
5386 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5389 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5390 command called pgrep for grepping around the active processes.
5392 7. Added the beginnings of support for UTF-8 character strings.
5394 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5395 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5396 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5397 out for the ar command.)
5400 Version 3.2 12-May-00
5401 ---------------------
5403 This is purely a bug fixing release.
5405 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5406 of ZA. This was just one example of several cases that could provoke this bug,
5407 which was introduced by change 9 of version 2.00. The code for breaking
5408 infinite loops after an iteration that matches an empty string was't working
5411 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5412 when matched against abbab (for example). After matching an empty string, it
5413 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5414 caused it to match further down the string than it should.
5416 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5417 was there because it doesn't seem to be needed, and it causes trouble on some
5418 systems, as it is not a Standard C header. It has been removed.
5420 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5421 were reported on the Macintosh. The changes were from
5423 while ((c = *(++ptr)) != 0 && c != '\n');
5425 while ((c = *(++ptr)) != 0 && c != '\n') ;
5427 Totally extraordinary, but if that's what it takes...
5429 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5430 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5431 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5432 assumes the way PCRE uses memmove() (always moving upwards).
5434 6. PCRE is being used in one environment where strchr() is not available. There
5435 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5439 Version 3.1 09-Feb-00
5440 ---------------------
5442 The only change in this release is the fixing of some bugs in Makefile.in for
5443 the "install" target:
5445 (1) It was failing to install pcreposix.h.
5447 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5450 Version 3.0 01-Feb-00
5451 ---------------------
5453 1. Add support for the /+ modifier to perltest (to output $` like it does in
5456 2. Add support for the /g modifier to perltest.
5458 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5459 matches null strings.
5461 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5462 pattern. Perl treats empty patterns specially - it reuses the most recent
5463 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5466 5. The POSIX interface was broken in that it was just handing over the POSIX
5467 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5468 required a bigger vector, with some working space on the end. This means that
5469 the POSIX wrapper now has to get and free some memory, and copy the results.
5471 6. Added some simple autoconf support, placing the test data and the
5472 documentation in separate directories, re-organizing some of the
5473 information files, and making it build pcre-config (a GNU standard). Also added
5474 libtool support for building PCRE as a shared library, which is now the
5477 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5478 09 are not valid octal constants. Single digits will be used for minor values
5481 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5482 existing programs that set these in the POSIX interface can use PCRE without
5485 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5486 return all that pcre_info() returns, plus additional data. The pcre_info()
5487 function is retained for compatibility, but is considered to be obsolete.
5489 10. Added experimental recursion feature (?R) to handle one common case that
5490 Perl 5.6 will be able to do with (?p{...}).
5492 11. Added support for POSIX character classes like [:alpha:], which Perl is
5496 Version 2.08 31-Aug-99
5497 ----------------------
5499 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5500 trying to match at the startoffset position, but instead was moving forward to
5501 the next newline as if a previous match had failed.
5503 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5504 and could get into a loop if a null string was matched other than at the start
5507 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5508 be distinguished at compile time, and for completeness also added PCRE_DATE.
5510 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5511 in GnuWin32 environments.
5514 Version 2.07 29-Jul-99
5515 ----------------------
5517 1. The documentation is now supplied in plain text form and HTML as well as in
5518 the form of man page sources.
5520 2. C++ compilers don't like assigning (void *) values to other pointer types.
5521 In particular this affects malloc(). Although there is no problem in Standard
5522 C, I've put in casts to keep C++ compilers happy.
5524 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5525 should be (const char *).
5527 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5528 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5529 However, I haven't made this a standard facility. The documentation doesn't
5530 mention it, and the Makefile doesn't support it.
5532 5. The Makefile now contains an "install" target, with editable destinations at
5533 the top of the file. The pcretest program is not installed.
5535 6. pgrep -V now gives the PCRE version number and date.
5537 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5538 causing the entire string to be ignored, instead of just the last character.
5540 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5541 non-matching string, it can take a very, very long time, even for strings of
5542 quite modest length, because of the nested recursion. PCRE now does better in
5543 some of these cases. It does this by remembering the last required literal
5544 character in the pattern, and pre-searching the subject to ensure it is present
5545 before running the real match. In other words, it applies a heuristic to detect
5546 some types of certain failure quickly, and in the above example, if presented
5547 with a string that has no trailing " it gives "no match" very quickly.
5549 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5550 other alternatives are tried instead.
5553 Version 2.06 09-Jun-99
5554 ----------------------
5556 1. Change pcretest's output for amount of store used to show just the code
5557 space, because the remainder (the data block) varies in size between 32-bit and
5560 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5561 start matching at. This allows lookbehinds to work when searching for multiple
5562 occurrences in a string.
5564 3. Added additional options to pcretest for testing multiple occurrences:
5566 /+ outputs the rest of the string that follows a match
5567 /g loops for multiple occurrences, using the new startoffset argument
5568 /G loops for multiple occurrences by passing an incremented pointer
5570 4. PCRE wasn't doing the "first character" optimization for patterns starting
5571 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5572 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5573 the letter 'x'. On long subject strings, this gives a significant speed-up.
5576 Version 2.05 21-Apr-99
5577 ----------------------
5579 1. Changed the type of magic_number from int to long int so that it works
5580 properly on 16-bit systems.
5582 2. Fixed a bug which caused patterns starting with .* not to work correctly
5583 when the subject string contained newline characters. PCRE was assuming
5584 anchoring for such patterns in all cases, which is not correct because .* will
5585 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5586 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5587 must be retried after every newline in the subject.
5590 Version 2.04 18-Feb-99
5591 ----------------------
5593 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5594 computation of the store needed to hold the pattern was incorrect (too large).
5595 If such patterns were nested a few deep, this could multiply and become a real
5598 2. Added /M option to pcretest to show the memory requirement of a specific
5599 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5601 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5602 compiled in such a way that the backtracking after subsequent failure was
5603 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5604 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5607 Version 2.03 02-Feb-99
5608 ----------------------
5610 1. Fixed typo and small mistake in man page.
5612 2. Added 4th condition (GPL supersedes if conflict) and created separate
5613 LICENCE file containing the conditions.
5615 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5616 Perl, that is the internal \ allows the delimiter to be included in the
5617 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5618 the final delimiter, add \ to the end of the pattern (to test the error).
5620 4. Added the convenience functions for extracting substrings after a successful
5621 match. Updated pcretest to make it able to test these functions.
5624 Version 2.02 14-Jan-99
5625 ----------------------
5627 1. Initialized the working variables associated with each extraction so that
5628 their saving and restoring doesn't refer to uninitialized store.
5630 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5631 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5634 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5635 calls, and wasn't printing the correct value for compiling calls. Increased the
5636 default value of LOOPREPEAT, and the number of significant figures in the
5639 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5641 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5642 a building problem on Windows NT with a FAT file system.
5645 Version 2.01 21-Oct-98
5646 ----------------------
5648 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5649 to character tables built by pcre_maketables() in the current locale. If NULL
5650 is passed, the default tables are used.
5653 Version 2.00 24-Sep-98
5654 ----------------------
5656 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5659 2. Allow quantification of (?>) groups, and make it work correctly.
5661 3. The first character computation wasn't working for (?>) groups.
5663 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5664 end of the subject) and add 5.005's \z, which really does match only at the
5665 very end of the subject.
5667 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5669 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5670 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5671 localized options. All options to pcre_study() were also removed.
5673 7. Add other new features from 5.005:
5675 $(?<= positive lookbehind
5676 $(?<! negative lookbehind
5677 (?imsx-imsx) added the unsetting capability
5678 such a setting is global if at outer level; local otherwise
5679 (?imsx-imsx:) non-capturing groups with option setting
5680 (?(cond)re|re) conditional pattern matching
5682 A backreference to itself in a repeated group matches the previous
5685 8. General tidying up of studying (both automatic and via "study")
5686 consequential on the addition of new assertions.
5688 9. As in 5.005, unlimited repeated groups that could match an empty substring
5689 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5690 runtime if any iteration does actually match an empty substring.
5692 10. Include the RunTest script in the distribution.
5694 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5695 discrepancies, some of which were old and were also with respect to 5.004. They
5696 have now been fixed.
5699 Version 1.09 28-Apr-98
5700 ----------------------
5702 1. A negated single character class followed by a quantifier with a minimum
5703 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5704 program crashes, or just wrong answers. This did not apply to negated classes
5705 containing more than one character, or to minima other than one.
5708 Version 1.08 27-Mar-98
5709 ----------------------
5711 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5713 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5714 latter must appear before anything that relies on it in the pattern.
5717 Version 1.07 16-Feb-98
5718 ----------------------
5720 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5721 repeat of a potentially empty string).
5724 Version 1.06 23-Jan-98
5725 ----------------------
5727 1. Added Markus Oberhumer's little patches for C++.
5729 2. Literal strings longer than 255 characters were broken.
5732 Version 1.05 23-Dec-97
5733 ----------------------
5735 1. Negated character classes containing more than one character were failing if
5736 PCRE_CASELESS was set at run time.
5739 Version 1.04 19-Dec-97
5740 ----------------------
5742 1. Corrected the man page, where some "const" qualifiers had been omitted.
5744 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5747 3. Fixed memory leak which occurred when a regex with back references was
5748 matched with an offsets vector that wasn't big enough. The temporary memory
5749 that is used in this case wasn't being freed if the match failed.
5751 4. Tidied pcretest to ensure it frees memory that it gets.
5753 5. Temporary memory was being obtained in the case where the passed offsets
5754 vector was exactly big enough.
5756 6. Corrected definition of offsetof() from change 5 below.
5758 7. I had screwed up change 6 below and broken the rules for the use of
5759 setjmp(). Now fixed.
5762 Version 1.03 18-Dec-97
5763 ----------------------
5765 1. A erroneous regex with a missing opening parenthesis was correctly
5766 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5769 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5770 it was reported that one broken compiler failed on the former because "code" is
5771 also an independent variable.
5773 3. The erroneous regex a[]b caused an array overrun reference.
5775 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5776 fail on data ending with that character. (It was going on too far, and checking
5777 the next character, typically a binary zero.) This was specific to the
5778 optimized code for single-character negative classes.
5780 5. Added a contributed patch from the TIN world which does the following:
5782 + Add an undef for memmove, in case the the system defines a macro for it.
5784 + Add a definition of offsetof(), in case there isn't one. (I don't know
5785 the reason behind this - offsetof() is part of the ANSI standard - but
5788 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5789 most of the places where whitespace preceded '#'. I have given up and
5790 allowed the remaining 2 cases to be at the margin.
5792 + Rename some variables in pcre to eliminate shadowing. This seems very
5793 pedantic, but does no harm, of course.
5795 6. Moved the call to setjmp() into its own function, to get rid of warnings
5796 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5798 7. Constructs such as \d{8,} were compiling into the equivalent of
5799 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5800 outcome, but in this particular case used more store than had been allocated,
5801 which caused the bug to be discovered because it threw up an internal error.
5803 8. The debugging code in both pcre and pcretest for outputting the compiled
5804 form of a regex was going wrong in the case of back references followed by
5805 curly-bracketed repeats.
5808 Version 1.02 12-Dec-97
5809 ----------------------
5811 1. Typos in pcre.3 and comments in the source fixed.
5813 2. Applied a contributed patch to get rid of places where it used to remove
5814 'const' from variables, and fixed some signed/unsigned and uninitialized
5817 3. Added the "runtest" target to Makefile.
5819 4. Set default compiler flag to -O2 rather than just -O.
5822 Version 1.01 19-Nov-97
5823 ----------------------
5825 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5826 like /([ab]*)*/, that is, for classes with more than one character in them.
5828 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5829 as /((?>a*))*/ (a PCRE_EXTRA facility).
5832 Version 1.00 18-Nov-97
5833 ----------------------
5835 1. Added compile-time macros to support systems such as SunOS4 which don't have
5836 memmove() or strerror() but have other things that can be used instead.
5838 2. Arranged that "make clean" removes the executables.
5841 Version 0.99 27-Oct-97
5842 ----------------------
5844 1. Fixed bug in code for optimizing classes with only one character. It was
5845 initializing a 32-byte map regardless, which could cause it to run off the end
5846 of the memory it had got.
5848 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5851 Version 0.98 22-Oct-97
5852 ----------------------
5854 1. Fixed bug in code for handling temporary memory usage when there are more
5855 back references than supplied space in the ovector. This could cause segfaults.
5858 Version 0.97 21-Oct-97
5859 ----------------------
5861 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5863 2. Optimized negated single characters not to use a bit map.
5865 3. Brought error texts together as macro definitions; clarified some of them;
5866 fixed one that was wrong - it said "range out of order" when it meant "invalid
5869 4. Changed some char * arguments to const char *.
5871 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5873 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5877 Version 0.96 16-Oct-97
5878 ----------------------
5880 1. Added a simple "pgrep" utility to the distribution.
5882 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5883 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5884 where "ddd" means "one or more decimal digits".
5886 3. Fixed serious bug. If a pattern had a back reference, but the call to
5887 pcre_exec() didn't supply a large enough ovector to record the related
5888 identifying subpattern, the match always failed. PCRE now remembers the number
5889 of the largest back reference, and gets some temporary memory in which to save
5890 the offsets during matching if necessary, in order to ensure that
5891 backreferences always work.
5893 4. Increased the compatibility with Perl in a number of ways:
5895 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5896 to request this handling. The option can be set at compile or exec time.
5898 (b) $ matches before a terminating newline by default; an option
5899 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5900 mode). The option can be set at compile or exec time.
5902 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5903 the same as Perl's. If the decimal number it represents is less than 10
5904 or there aren't that many previous left capturing parentheses, an octal
5905 escape is read. Inside a character class, it's always an octal escape,
5906 even if it is a single digit.
5908 (d) An escaped but undefined alphabetic character is taken as a literal,
5909 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5912 (e) {0} is now permitted. (The previous item is removed from the compiled
5915 5. Changed all the names of code files so that the basic parts are no longer
5916 than 10 characters, and abolished the teeny "globals.c" file.
5918 6. Changed the handling of character classes; they are now done with a 32-byte
5921 7. Added the -d and /D options to pcretest to make it possible to look at the
5922 internals of compilation without having to recompile pcre.
5925 Version 0.95 23-Sep-97
5926 ----------------------
5928 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5929 \x20 at the start of a run of normal characters. These were being treated as
5930 real characters, instead of the source characters being re-checked.
5933 Version 0.94 18-Sep-97
5934 ----------------------
5936 1. The functions are now thread-safe, with the caveat that the global variables
5937 containing pointers to malloc() and free() or alternative functions are the
5938 same for all threads.
5940 2. Get pcre_study() to generate a bitmap of initial characters for non-
5941 anchored patterns when this is possible, and use it if passed to pcre_exec().
5944 Version 0.93 15-Sep-97
5945 ----------------------
5947 1. /(b)|(:+)/ was computing an incorrect first character.
5949 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5950 but not actually doing anything yet.
5952 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5953 as Perl does (e.g. [-az] or [az-]).
5955 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5956 all possible positions.
5958 5. Split up into different modules to avoid including unneeded functions in a
5959 compiled binary. However, compile and exec are still in one module. The "study"
5960 function is split off.
5962 6. The character tables are now in a separate module whose source is generated
5963 by an auxiliary program - but can then be edited by hand if required. There are
5964 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5965 toupper() in the code.
5967 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5968 make them global. Abolish the function for setting them, as the caller can now
5972 Version 0.92 11-Sep-97
5973 ----------------------
5975 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5976 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5978 2. Caseless matching was not working in character classes if the characters in
5979 the pattern were in upper case.
5981 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5983 4. Make PCRE_ANCHORED public and accept as a compile option.
5985 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5986 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5989 6. Give an error if bad option bits passed at compile or run time.
5991 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5992 pcretest to cause it to pass that flag.
5994 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5995 options, and the first character, if set.
5997 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
6000 Version 0.91 10-Sep-97
6001 ----------------------
6003 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
6004 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
6006 2. PCRE was looping on encountering an indefinitely repeated back reference to
6007 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
6008 Perl does - treats the match as successful.