ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1623 - (show annotations)
Thu Feb 4 17:13:10 2016 UTC (3 years, 7 months ago) by zherczeg
File size: 287476 byte(s)
JIT code generator for assertion matching is refactored to a separat function.
1 ChangeLog for PCRE
2 ------------------
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
7 Version 8.39 xx-xxxxxx-201x
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. An call of pcre_copy_named_substring() for a named substring whose number
37 was greater than the space in the ovector could cause a crash.
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. JIT code generator for assertion matching is refactored to a separate
48 function.
51 Version 8.38 23-November-2015
52 -----------------------------
54 1. If a group that contained a recursive back reference also contained a
55 forward reference subroutine call followed by a non-forward-reference
56 subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
57 compile correct code, leading to undefined behaviour or an internally
58 detected error. This bug was discovered by the LLVM fuzzer.
60 2. Quantification of certain items (e.g. atomic back references) could cause
61 incorrect code to be compiled when recursive forward references were
62 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
63 This bug was discovered by the LLVM fuzzer.
65 3. A repeated conditional group whose condition was a reference by name caused
66 a buffer overflow if there was more than one group with the given name.
67 This bug was discovered by the LLVM fuzzer.
69 4. A recursive back reference by name within a group that had the same name as
70 another group caused a buffer overflow. For example:
71 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
73 5. A forward reference by name to a group whose number is the same as the
74 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
75 a buffer overflow at compile time. This bug was discovered by the LLVM
76 fuzzer.
78 6. A lookbehind assertion within a set of mutually recursive subpatterns could
79 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
81 7. Another buffer overflow bug involved duplicate named groups with a
82 reference between their definition, with a group that reset capture
83 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
84 fixed by always allowing for more memory, even if not needed. (A proper fix
85 is implemented in PCRE2, but it involves more refactoring.)
87 8. There was no check for integer overflow in subroutine calls such as (?123).
89 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
90 being treated as a literal 'l' instead of causing an error.
92 10. There was a buffer overflow if pcre_exec() was called with an ovector of
93 size 1. This bug was found by american fuzzy lop.
95 11. If a non-capturing group containing a conditional group that could match
96 an empty string was repeated, it was not identified as matching an empty
97 string itself. For example: /^(?:(?(1)x|)+)+$()/.
99 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
100 \a and \e in test subject lines.
102 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
103 instead of the EBCDIC value.
105 14. The handling of \c in an EBCDIC environment has been revised so that it is
106 now compatible with the specification in Perl's perlebcdic page.
108 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
109 ASCII/Unicode. This has now been added to the list of characters that are
110 recognized as white space in EBCDIC.
112 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
113 error (correctly) when used outside a class, but did not give an error
114 within a class.
116 17. \h within a class was incorrectly compiled in EBCDIC environments.
118 18. A pattern with an unmatched closing parenthesis that contained a backward
119 assertion which itself contained a forward reference caused buffer
120 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
122 19. JIT should return with error when the compiled pattern requires more stack
123 space than the maximum.
125 20. A possessively repeated conditional group that could match an empty string,
126 for example, /(?(R))*+/, was incorrectly compiled.
128 21. Fix infinite recursion in the JIT compiler when certain patterns such as
129 /(?:|a|){100}x/ are analysed.
131 22. Some patterns with character classes involving [: and \\ were incorrectly
132 compiled and could cause reading from uninitialized memory or an incorrect
133 error diagnosis.
135 23. Pathological patterns containing many nested occurrences of [: caused
136 pcre_compile() to run for a very long time.
138 24. A conditional group with only one branch has an implicit empty alternative
139 branch and must therefore be treated as potentially matching an empty
140 string.
142 25. If (?R was followed by - or + incorrect behaviour happened instead of a
143 diagnostic.
145 26. Arrange to give up on finding the minimum matching length for overly
146 complex patterns.
148 27. Similar to (4) above: in a pattern with duplicated named groups and an
149 occurrence of (?| it is possible for an apparently non-recursive back
150 reference to become recursive if a later named group with the relevant
151 number is encountered. This could lead to a buffer overflow. Wen Guanxing
152 from Venustech ADLAB discovered this bug.
154 28. If pcregrep was given the -q option with -c or -l, or when handling a
155 binary file, it incorrectly wrote output to stdout.
157 29. The JIT compiler did not restore the control verb head in case of *THEN
158 control verbs. This issue was found by Karl Skomski with a custom LLVM
159 fuzzer.
161 30. Error messages for syntax errors following \g and \k were giving inaccurate
162 offsets in the pattern.
164 31. Added a check for integer overflow in conditions (?(<digits>) and
165 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
166 fuzzer.
168 32. Handling recursive references such as (?2) when the reference is to a group
169 later in the pattern uses code that is very hacked about and error-prone.
170 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
171 give an internal error if it is obvious that compiling has gone wrong.
173 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
174 This issue was found by Karl Skomski with a custom LLVM fuzzer.
176 34. The JIT compiler should restore the control chain for empty possessive
177 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
179 35. Match limit check added to JIT recursion. This issue was found by Karl
180 Skomski with a custom LLVM fuzzer.
182 36. Yet another case similar to 27 above has been circumvented by an
183 unconditional allocation of extra memory. This issue is fixed "properly" in
184 PCRE2 by refactoring the way references are handled. Wen Guanxing
185 from Venustech ADLAB discovered this bug.
187 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
188 with a custom LLVM fuzzer.
190 38. Fixed a corner case of range optimization in JIT.
192 39. An incorrect error "overran compiling workspace" was given if there were
193 exactly enough group forward references such that the last one extended
194 into the workspace safety margin. The next one would have expanded the
195 workspace. The test for overflow was not including the safety margin.
197 40. A match limit issue is fixed in JIT which was found by Karl Skomski
198 with a custom LLVM fuzzer.
200 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
201 work under Windows. (Why has it taken so long for anyone to notice?)
203 42. In a character class such as [\W\p{Any}] where both a negative-type escape
204 ("not a word character") and a property escape were present, the property
205 escape was being ignored.
207 43. Fix crash caused by very long (*MARK) or (*THEN) names.
209 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
210 by a single ASCII character in a class item, was incorrectly compiled in
211 UCP mode. The POSIX class got lost, but only if the single character
212 followed it.
214 45. [:punct:] in UCP mode was matching some characters in the range 128-255
215 that should not have been matched.
217 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
218 class, all characters with code points greater than 255 are in the class.
219 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
220 such as \w are turned into Unicode properties), wide characters were not
221 correctly handled, and could fail to match.
224 Version 8.37 28-April-2015
225 --------------------------
227 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
228 for those parentheses to be closed with whatever has been captured so far.
229 However, it was failing to mark any other groups between the hightest
230 capture so far and the currrent group as "unset". Thus, the ovector for
231 those groups contained whatever was previously there. An example is the
232 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
234 2. If an assertion condition was quantified with a minimum of zero (an odd
235 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
237 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
238 unrecognized modifier, a crash could occur.
240 4. An attempt to do global matching in pcretest with a zero-length ovector
241 caused a crash.
243 5. Fixed a memory leak during matching that could occur for a subpattern
244 subroutine call (recursive or otherwise) if the number of captured groups
245 that had to be saved was greater than ten.
247 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
248 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
249 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
251 7. A UTF pattern containing a "not" match of a non-ASCII character and a
252 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
254 8. When a pattern is compiled, it remembers the highest back reference so that
255 when matching, if the ovector is too small, extra memory can be obtained to
256 use instead. A conditional subpattern whose condition is a check on a
257 capture having happened, such as, for example in the pattern
258 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
259 setting the highest backreference number. This mattered only if pcre_exec()
260 was called with an ovector that was too small to hold the capture, and there
261 was no other kind of back reference (a situation which is probably quite
262 rare). The effect of the bug was that the condition was always treated as
263 FALSE when the capture could not be consulted, leading to a incorrect
264 behaviour by pcre_exec(). This bug has been fixed.
266 9. A reference to a duplicated named group (either a back reference or a test
267 for being set in a conditional) that occurred in a part of the pattern where
268 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
269 to be incorrectly calculated, leading to overwriting.
271 10. A mutually recursive set of back references such as (\2)(\1) caused a
272 segfault at study time (while trying to find the minimum matching length).
273 The infinite loop is now broken (with the minimum length unset, that is,
274 zero).
276 11. If an assertion that was used as a condition was quantified with a minimum
277 of zero, matching went wrong. In particular, if the whole group had
278 unlimited repetition and could match an empty string, a segfault was
279 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
280 assertions to be quantified, but not if they are being used as conditions,
281 so the above pattern is faulted by Perl. PCRE has now been changed so that
282 it also rejects such patterns.
284 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
285 failed to allow the zero-repeat case if pcre2_exec() was called with an
286 ovector too small to capture the group.
288 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
289 Red Hat Product Security:
291 (a) A crash if /K and /F were both set with the option to save the compiled
292 pattern.
294 (b) Another crash if the option to print captured substrings in a callout
295 was combined with setting a null ovector, for example \O\C+ as a subject
296 string.
298 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
299 forward reference repeated a large (but limited) number of times within a
300 repeated outer group that has a zero minimum quantifier, caused incorrect
301 code to be compiled, leading to the error "internal error:
302 previously-checked referenced subpattern not found" when an incorrect
303 memory address was read. This bug was reported as "heap overflow",
304 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
305 CVE-2015-2325.
307 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
308 call within a group that also contained a recursive back reference caused
309 incorrect code to be compiled. This bug was reported as "heap overflow",
310 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
311 number CVE-2015-2326.
313 24. Computing the size of the JIT read-only data in advance has been a source
314 of various issues, and new ones are still appear unfortunately. To fix
315 existing and future issues, size computation is eliminated from the code,
316 and replaced by on-demand memory allocation.
318 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
319 adjacent to the end of the range, and the range contained characters with
320 more than one other case, caused incorrect behaviour when compiled in UTF
321 mode. In that example, the range a-j was left out of the class.
323 26. Fix JIT compilation of conditional blocks, which assertion
324 is converted to (*FAIL). E.g: /(?(?!))/.
326 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
327 discovered by the LLVM fuzzer.
329 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
330 when this assertion was used as a condition, for example (?(?!)a|b). In
331 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
332 error about an unsupported item.
334 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
335 possessification code could take exponential time to complete. A recursion
336 depth limit of 1000 has been imposed to limit the resources used by this
337 optimization.
339 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
340 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
341 because \S ensures they are all in the class. The code for doing this was
342 interacting badly with the code for computing the amount of space needed to
343 compile the pattern, leading to a buffer overflow. This bug was discovered
344 by the LLVM fuzzer.
346 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
347 other kinds of group caused stack overflow at compile time. This bug was
348 discovered by the LLVM fuzzer.
350 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
351 between a subroutine call and its quantifier was incorrectly compiled,
352 leading to buffer overflow or other errors. This bug was discovered by the
353 LLVM fuzzer.
355 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
356 assertion after (?(. The code was failing to check the character after
357 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
358 was discovered by the LLVM fuzzer.
360 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
361 a fixed maximum following a group that contains a subroutine reference was
362 incorrectly compiled and could trigger buffer overflow. This bug was
363 discovered by the LLVM fuzzer.
365 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
366 caused a stack overflow instead of the diagnosis of a non-fixed length
367 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
369 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
370 (e.g. /(?<=\Ka)/) could make pcregrep loop.
372 37. There was a similar problem to 36 in pcretest for global matches.
374 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
375 and a subsequent item in the pattern caused a non-match, backtracking over
376 the repeated \X did not stop, but carried on past the start of the subject,
377 causing reference to random memory and/or a segfault. There were also some
378 other cases where backtracking after \C could crash. This set of bugs was
379 discovered by the LLVM fuzzer.
381 39. The function for finding the minimum length of a matching string could take
382 a very long time if mutual recursion was present many times in a pattern,
383 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
384 method has been implemented. This infelicity was discovered by the LLVM
385 fuzzer.
387 40. Static linking against the PCRE library using the pkg-config module was
388 failing on missing pthread symbols.
391 Version 8.36 26-September-2014
392 ------------------------------
394 1. Got rid of some compiler warnings in the C++ modules that were shown up by
395 -Wmissing-field-initializers and -Wunused-parameter.
397 2. The tests for quantifiers being too big (greater than 65535) were being
398 applied after reading the number, and stupidly assuming that integer
399 overflow would give a negative number. The tests are now applied as the
400 numbers are read.
402 3. Tidy code in pcre_exec.c where two branches that used to be different are
403 now the same.
405 4. The JIT compiler did not generate match limit checks for certain
406 bracketed expressions with quantifiers. This may lead to exponential
407 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
408 issue should be resolved now.
410 5. Fixed an issue, which occures when nested alternatives are optimized
411 with table jumps.
413 6. Inserted two casts and changed some ints to size_t in the light of some
414 reported 64-bit compiler warnings (Bugzilla 1477).
416 7. Fixed a bug concerned with zero-minimum possessive groups that could match
417 an empty string, which sometimes were behaving incorrectly in the
418 interpreter (though correctly in the JIT matcher). This pcretest input is
419 an example:
421 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
424 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
425 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
426 for an empty string was breaking the inner loop and carrying on at a lower
427 level, when possessive repeated groups should always return to a higher
428 level as they have no backtrack points in them. The empty string test now
429 occurs at the outer level.
431 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
432 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
434 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
435 doesn't).
437 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
438 to hold a pattern to be incorrectly computed (too small) when there were
439 named back references to duplicated names. This could cause "internal
440 error: code overflow" or "double free or corruption" or other memory
441 handling errors.
443 11. When named subpatterns had the same prefixes, back references could be
444 confused. For example, in this pattern:
446 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
448 the reference to 'Name' was incorrectly treated as a reference to a
449 duplicate name.
451 12. A pattern such as /^s?c/mi8 where the optional character has more than
452 one "other case" was incorrectly compiled such that it would only try to
453 match starting at "c".
455 13. When a pattern starting with \s was studied, VT was not included in the
456 list of possible starting characters; this should have been part of the
457 8.34/18 patch.
459 14. If a character class started [\Qx]... where x is any character, the class
460 was incorrectly terminated at the ].
462 15. If a pattern that started with a caseless match for a character with more
463 than one "other case" was studied, PCRE did not set up the starting code
464 unit bit map for the list of possible characters. Now it does. This is an
465 optimization improvement, not a bug fix.
467 16. The Unicode data tables have been updated to Unicode 7.0.0.
469 17. Fixed a number of memory leaks in pcregrep.
471 18. Avoid a compiler warning (from some compilers) for a function call with
472 a cast that removes "const" from an lvalue by using an intermediate
473 variable (to which the compiler does not object).
475 19. Incorrect code was compiled if a group that contained an internal recursive
476 back reference was optional (had quantifier with a minimum of zero). This
477 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
478 caused segmentation faults because of stack overflows at compile time.
480 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
481 group that is quantified with an indefinite repeat, caused a compile-time
482 loop which used up all the system stack and provoked a segmentation fault.
483 This was not the same bug as 19 above.
485 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
486 Patch by Mike Frysinger.
489 Version 8.35 04-April-2014
490 --------------------------
492 1. A new flag is set, when property checks are present in an XCLASS.
493 When this flag is not set, PCRE can perform certain optimizations
494 such as studying these XCLASS-es.
496 2. The auto-possessification of character sets were improved: a normal
497 and an extended character set can be compared now. Furthermore
498 the JIT compiler optimizes more character set checks.
500 3. Got rid of some compiler warnings for potentially uninitialized variables
501 that show up only when compiled with -O2.
503 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
504 of a match later then the end of the match. The pcretest program was not
505 handling the case sensibly - it was outputting from the start to the next
506 binary zero. It now reports this situation in a message, and outputs the
507 text from the end to the start.
509 5. Fast forward search is improved in JIT. Instead of the first three
510 characters, any three characters with fixed position can be searched.
511 Search order: first, last, middle.
513 6. Improve character range checks in JIT. Characters are read by an inprecise
514 function now, which returns with an unknown value if the character code is
515 above a certain threshold (e.g: 256). The only limitation is that the value
516 must be bigger than the threshold as well. This function is useful when
517 the characters above the threshold are handled in the same way.
519 7. The macros whose names start with RAWUCHAR are placeholders for a future
520 mode in which only the bottom 21 bits of 32-bit data items are used. To
521 make this more memorable for those maintaining the code, the names have
522 been changed to start with UCHAR21, and an extensive comment has been added
523 to their definition.
525 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
526 to the export list in Makefile.am (they were accidentally omitted from the
527 8.34 tarball).
529 9. The informational output from pcretest used the phrase "starting byte set"
530 which is inappropriate for the 16-bit and 32-bit libraries. As the output
531 for "first char" and "need char" really means "non-UTF-char", I've changed
532 "byte" to "char", and slightly reworded the output. The documentation about
533 these values has also been (I hope) clarified.
535 10. Another JIT related optimization: use table jumps for selecting the correct
536 backtracking path, when more than four alternatives are present inside a
537 bracket.
539 11. Empty match is not possible, when the minimum length is greater than zero,
540 and there is no \K in the pattern. JIT should avoid empty match checks in
541 such cases.
543 12. In a caseless character class with UCP support, when a character with more
544 than one alternative case was not the first character of a range, not all
545 the alternative cases were added to the class. For example, s and \x{17f}
546 are both alternative cases for S: the class [RST] was handled correctly,
547 but [R-T] was not.
549 13. The configure.ac file always checked for pthread support when JIT was
550 enabled. This is not used in Windows, so I have put this test inside a
551 check for the presence of windows.h (which was already tested for).
553 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
554 The algorithm provides a way to skip certain starting offsets, and usually
555 faster than linear prefix searches.
557 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
558 as for 'fr_FR' and 'french'. For some reason, however, it then used the
559 Windows-specific input and output files, which have 'french' screwed in.
560 So this could never have worked. One of the problems with locales is that
561 they aren't always the same. I have now updated RunTest so that it checks
562 the output of the locale test (test 3) against three different output
563 files, and it allows the test to pass if any one of them matches. With luck
564 this should make the test pass on some versions of Solaris where it was
565 failing. Because of the uncertainty, the script did not used to stop if
566 test 3 failed; it now does. If further versions of a French locale ever
567 come to light, they can now easily be added.
569 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
570 there was a message during ./configure, but it did not stop. This now
571 provokes an error. The invalid example in README has been corrected.
572 If a value less than the minimum is given, the minimum value has always
573 been used, but now a warning is given.
575 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
576 was a bug in the test system, which is now fixed. Also, the list of various
577 configurations that are tested for each release did not have one with both
578 16/32 bits and --enable-bar-anycrlf. It now does.
580 18. pcretest was missing "-C bsr" for displaying the \R default setting.
582 19. Little endian PowerPC systems are supported now by the JIT compiler.
584 20. The fast forward newline mechanism could enter to an infinite loop on
585 certain invalid UTF-8 input. Although we don't support these cases
586 this issue can be fixed by a performance optimization.
588 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
589 not take account if existing stack usage. There is now a new global
590 variable called pcre_stack_guard that can be set to point to an external
591 function to check stack availability. It is called at the start of
592 processing every parenthesized group.
594 22. A typo in the code meant that in ungreedy mode the max/min qualifier
595 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
596 not match "ab".
598 23. When UTF was disabled, the JIT program reported some incorrect compile
599 errors. These messages are silenced now.
601 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
602 compiler.
604 25. Change all the temporary files used in RunGrepTest to be different to those
605 used by RunTest so that the tests can be run simultaneously, for example by
606 "make -j check".
609 Version 8.34 15-December-2013
610 -----------------------------
612 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
613 executable memory. Patch inspired by Carsten Klein.
615 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
616 this macro is never tested and has no effect, because the work to support
617 coverage involves only compiling and linking options and special targets in
618 the Makefile. The comment in config.h implied that defining the macro would
619 enable coverage support, which is totally false. There was also support for
620 setting this macro in the CMake files (my fault, I just copied it from
621 configure). SUPPORT_GCOV has now been removed.
623 3. Make a small performance improvement in strlen16() and strlen32() in
624 pcretest.
626 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
627 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
628 these cases). There was also one in pcretest.c.
630 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
632 6. In UTF mode, the code for checking whether a group could match an empty
633 string (which is used for indefinitely repeated groups to allow for
634 breaking an infinite loop) was broken when the group contained a repeated
635 negated single-character class with a character that occupied more than one
636 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
637 UTF-8 mode). The effect was undefined: the group might or might not be
638 deemed as matching an empty string, or the program might have crashed.
640 7. The code for checking whether a group could match an empty string was not
641 recognizing that \h, \H, \v, \V, and \R must match a character.
643 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
644 an empty string. If it can, pcretest shows this in its information output.
646 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
647 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
648 matched by pcre_exec() without using JIT:
650 (a) If the rest of the pattern did not match after a maximal run of
651 grapheme clusters, the code for backing up to try with fewer of them
652 did not always back up over a full grapheme when characters that do not
653 have the modifier quality were involved, e.g. Hangul syllables.
655 (b) If the match point in a subject started with modifier character, and
656 there was no match, the code could incorrectly back up beyond the match
657 point, and potentially beyond the first character in the subject,
658 leading to a segfault or an incorrect match result.
660 10. A conditional group with an assertion condition could lead to PCRE
661 recording an incorrect first data item for a match if no other first data
662 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
663 first data item, and therefore matched "ca" after "c" instead of at the
664 start.
666 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
667 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
669 12. The source of pcregrep now includes z/OS-specific code so that it can be
670 compiled for z/OS as part of the special z/OS distribution.
672 13. Added the -T and -TM options to pcretest.
674 14. The code in pcre_compile.c for creating the table of named capturing groups
675 has been refactored. Instead of creating the table dynamically during the
676 actual compiling pass, the information is remembered during the pre-compile
677 pass (on the stack unless there are more than 20 named groups, in which
678 case malloc() is used) and the whole table is created before the actual
679 compile happens. This has simplified the code (it is now nearly 150 lines
680 shorter) and prepared the way for better handling of references to groups
681 with duplicate names.
683 15. A back reference to a named subpattern when there is more than one of the
684 same name now checks them in the order in which they appear in the pattern.
685 The first one that is set is used for the reference. Previously only the
686 first one was inspected. This change makes PCRE more compatible with Perl.
688 16. Unicode character properties were updated from Unicode 6.3.0.
690 17. The compile-time code for auto-possessification has been refactored, based
691 on a patch by Zoltan Herczeg. It now happens after instead of during
692 compilation. The code is cleaner, and more cases are handled. The option
693 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
694 options in pcretest are provided to set it. It can also be set by
695 (*NO_AUTO_POSSESS) at the start of a pattern.
697 18. The character VT has been added to the default ("C" locale) set of
698 characters that match \s and are generally treated as white space,
699 following this same change in Perl 5.18. There is now no difference between
700 "Perl space" and "POSIX space". Whether VT is treated as white space in
701 other locales depends on the locale.
703 19. The code for checking named groups as conditions, either for being set or
704 for being recursed, has been refactored (this is related to 14 and 15
705 above). Processing unduplicated named groups should now be as fast at
706 numerical groups, and processing duplicated groups should be faster than
707 before.
709 20. Two patches to the CMake build system, by Alexander Barkov:
711 (1) Replace the "source" command by "." in CMakeLists.txt because
712 "source" is a bash-ism.
714 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
715 without these the CMake build does not work on Solaris.
717 21. Perl has changed its handling of \8 and \9. If there is no previously
718 encountered capturing group of those numbers, they are treated as the
719 literal characters 8 and 9 instead of a binary zero followed by the
720 literals. PCRE now does the same.
722 22. Following Perl, added \o{} to specify codepoints in octal, making it
723 possible to specify values greater than 0777 and also making them
724 unambiguous.
726 23. Perl now gives an error for missing closing braces after \x{... instead of
727 treating the string as literal. PCRE now does the same.
729 24. RunTest used to grumble if an inappropriate test was selected explicitly,
730 but just skip it when running all tests. This make it awkward to run ranges
731 of tests when one of them was inappropriate. Now it just skips any
732 inappropriate tests, as it always did when running all tests.
734 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
735 character types such as \d or \w, too many callouts were inserted, and the
736 data that they returned was rubbish.
738 26. In UCP mode, \s was not matching two of the characters that Perl matches,
739 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
740 were matched by \h. The code has now been refactored so that the lists of
741 the horizontal and vertical whitespace characters used for \h and \v (which
742 are defined only in one place) are now also used for \s.
744 27. Add JIT support for the 64 bit TileGX architecture.
745 Patch by Jiong Wang (Tilera Corporation).
747 28. Possessive quantifiers for classes (both explicit and automatically
748 generated) now use special opcodes instead of wrapping in ONCE brackets.
750 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
751 (because it's meaningless), this was not happening when PCRE_CASELESS was
752 set. Not wrong, but inefficient.
754 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
755 properties for \w, \d, etc) is present in a test regex. Otherwise if the
756 test contains no characters greater than 255, Perl doesn't realise it
757 should be using Unicode semantics.
759 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
760 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
761 does in Unicode mode.
763 32. Added the "forbid" facility to pcretest so that putting tests into the
764 wrong test files can sometimes be quickly detected.
766 33. There is now a limit (default 250) on the depth of nesting of parentheses.
767 This limit is imposed to control the amount of system stack used at compile
768 time. It can be changed at build time by --with-parens-nest-limit=xxx or
769 the equivalent in CMake.
771 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
772 errors. Perl warns for these when in warning mode, but PCRE has no facility
773 for giving warnings.
775 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
776 However, this was not working for (?!) because it is optimized to (*FAIL),
777 for which PCRE does not allow quantifiers. The optimization is now disabled
778 when a quantifier follows (?!). I can't see any use for this, but it makes
779 things uniform.
781 36. Perl no longer allows group names to start with digits, so I have made this
782 change also in PCRE. It simplifies the code a bit.
784 37. In extended mode, Perl ignores spaces before a + that indicates a
785 possessive quantifier. PCRE allowed a space before the quantifier, but not
786 before the possessive +. It now does.
788 38. The use of \K (reset reported match start) within a repeated possessive
789 group such as (a\Kb)*+ was not working.
791 40. Document that the same character tables must be used at compile time and
792 run time, and that the facility to pass tables to pcre_exec() and
793 pcre_dfa_exec() is for use only with saved/restored patterns.
795 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
796 features for Builds with MSVC:
798 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
799 builds). The .rc files can be used to set FileDescription and many other
800 attributes.
802 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
803 This allows higher-level build scripts which want .pdb files to avoid
804 hard-coding the exact files needed."
806 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
807 mean "start of word" and "end of word", respectively, as a transition aid.
809 43. A minimizing repeat of a class containing codepoints greater than 255 in
810 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
811 compiled to use the heap for recursion.
813 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
814 is configured.
817 Version 8.33 28-May-2013
818 ------------------------
820 1. Added 'U' to some constants that are compared to unsigned integers, to
821 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
822 variables that are added to signed variables, to ensure the result is
823 signed and can be negated.
825 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
826 CMake config files.
828 3. Revise the creation of config.h.generic so that all boolean macros are
829 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
830 overriding via -D on the command line possible.
832 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
833 to unsigned int is reported to make a quite noticeable speed difference in
834 a specific Windows environment. Testing on Linux did also appear to show
835 some benefit (and it is clearly not harmful). Also fixed the definition of
836 Xop which should be unsigned.
838 5. Related to (4), changing the definition of the intermediate variable cc
839 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
840 improvements.
842 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
843 unnecessary spaces.
845 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
846 and later.
848 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
850 9. Optimizing fast_forward_start_bits in JIT.
852 10. Adding support for callouts in JIT, and fixing some issues revealed
853 during this work. Namely:
855 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
857 (b) Minimum length was not checked before the matching is started.
859 11. The value of capture_last that is passed to callouts was incorrect in some
860 cases when there was a capture on one path that was subsequently abandoned
861 after a backtrack. Also, the capture_last value is now reset after a
862 recursion, since all captures are also reset in this case.
864 12. The interpreter no longer returns the "too many substrings" error in the
865 case when an overflowing capture is in a branch that is subsequently
866 abandoned after a backtrack.
868 13. In the pathological case when an offset vector of size 2 is used, pcretest
869 now prints out the matched string after a yield of 0 or 1.
871 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
872 Only supported by the JIT compiler at the moment.
874 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
876 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
877 offset of the starting point of the matching process, provided the offsets
878 vector is large enough.
880 17. The \A escape now records a lookbehind value of 1, though its execution
881 does not actually inspect the previous character. This is to ensure that,
882 in partial multi-segment matching, at least one character from the old
883 segment is retained when a new segment is processed. Otherwise, if there
884 are no lookbehinds in the pattern, \A might match incorrectly at the start
885 of a new segment.
887 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
889 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
890 gives some modest performance improvement in 8-bit mode.
892 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
893 be expressed in certain programming languages using Universal Character
894 Names.
896 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
897 which points out that "non characters" are not "characters that may not
898 appear in Unicode strings" but rather "characters that are reserved for
899 internal use and have only local meaning".
901 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
902 there was a conditional group that depended on an assertion, if the
903 assertion was false, the callout that immediately followed the alternation
904 in the condition was skipped when pcre_exec() was used for matching.
906 23. Allow an explicit callout to be inserted before an assertion that is the
907 condition for a conditional group, for compatibility with automatic
908 callouts, which always insert a callout at this point.
910 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
911 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
913 25. (*PRUNE) is now supported by the JIT compiler.
915 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
917 27. Fix the case where there are two or more SKIPs with arguments that may be
918 ignored.
920 28. (*SKIP) is now supported by the JIT compiler.
922 29. (*THEN) is now supported by the JIT compiler.
924 30. Update RunTest with additional test selector options.
926 31. The way PCRE handles backtracking verbs has been changed in two ways.
928 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
929 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
930 backtracking. In some cases this makes it more Perl-compatible, but Perl's
931 rather obscure rules do not always do the same thing.
933 (2) Previously, backtracking verbs were confined within assertions. This is
934 no longer the case for positive assertions, except for (*ACCEPT). Again,
935 this sometimes improves Perl compatibility, and sometimes does not.
937 32. A number of tests that were in test 2 because Perl did things differently
938 have been moved to test 1, because either Perl or PCRE has changed, and
939 these tests are now compatible.
941 32. Backtracking control verbs are now handled in the same way in JIT and
942 interpreter.
944 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
945 contained a forward subroutine reference caused a compile error.
947 34. Auto-detect and optimize limited repetitions in JIT.
949 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
950 blocking (*UTF) etc.
952 36. In the interpreter, maximizing pattern repetitions for characters and
953 character types now use tail recursion, which reduces stack usage.
955 37. The value of the max lookbehind was not correctly preserved if a compiled
956 and saved regex was reloaded on a host of different endianness.
958 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
959 of the compiled pattern block, expand the flags field from 16 to 32 bits
960 because it was almost full.
962 39. Try madvise first before posix_madvise.
964 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
965 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
966 original change locked out all matches of empty strings. This has been
967 changed so that one match of an empty string per line is recognized.
968 Subsequent searches on the same line (for colouring or for --only-matching,
969 for example) do not recognize empty strings.
971 41. Applied a user patch to fix a number of spelling mistakes in comments.
973 42. Data lines longer than 65536 caused pcretest to crash.
975 43. Clarified the data type for length and startoffset arguments for pcre_exec
976 and pcre_dfa_exec in the function-specific man pages, where they were
977 explicitly stated to be in bytes, never having been updated. I also added
978 some clarification to the pcreapi man page.
980 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
981 a segmentation fault.
984 Version 8.32 30-November-2012
985 -----------------------------
987 1. Improved JIT compiler optimizations for first character search and single
988 character iterators.
990 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
991 Patch by Daniel Richard G.
993 3. Single character iterator optimizations in the JIT compiler.
995 4. Improved JIT compiler optimizations for character ranges.
997 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
998 Reported by Giuseppe D'Angelo.
1000 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1001 of a line, was being set incorrectly in cases where .* appeared inside
1002 atomic brackets at the start of a pattern, or where there was a subsequent
1003 *PRUNE or *SKIP.
1005 7. Improved instruction cache flush for POWER/PowerPC.
1006 Patch by Daniel Richard G.
1008 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
1009 grep:
1011 (a) There is now no limit to the number of patterns to be matched.
1013 (b) An error is given if a pattern is too long.
1015 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1016 are now supported.
1018 (d) --exclude-from and --include-from (multiple use) have been added.
1020 (e) Exclusions and inclusions now apply to all files and directories, not
1021 just to those obtained from scanning a directory recursively.
1023 (f) Multiple uses of -f and --file-list are now supported.
1025 (g) In a Windows environment, the default for -d has been changed from
1026 "read" (the GNU grep default) to "skip", because otherwise the presence
1027 of a directory in the file list provokes an error.
1029 (h) The documentation has been revised and clarified in places.
1031 9. Improve the matching speed of capturing brackets.
1033 10. Changed the meaning of \X so that it now matches a Unicode extended
1034 grapheme cluster.
1036 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1037 out POSIX threads when JIT support is configured.
1039 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1041 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1042 structure could go wrong in environments where size_t is not the same size
1043 as int.
1045 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1046 set.
1048 15. The EBCDIC support had decayed; later updates to the code had included
1049 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1050 general tidy up of EBCDIC-related issues, and the documentation was also
1051 not quite right. There is now a test that can be run on ASCII systems to
1052 check some of the EBCDIC-related things (but is it not a full test).
1054 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1055 in a small tidy to the code.
1057 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1059 18. If the --only-matching (-o) option in pcregrep is specified multiple
1060 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1061 substrings matched by the 1st and 2nd capturing parentheses. A separating
1062 string can be specified by --om-separator (default empty).
1064 19. Improving the first n character searches.
1066 20. Turn case lists for horizontal and vertical white space into macros so that
1067 they are defined only once.
1069 21. This set of changes together give more compatible Unicode case-folding
1070 behaviour for characters that have more than one other case when UCP
1071 support is available.
1073 (a) The Unicode property table now has offsets into a new table of sets of
1074 three or more characters that are case-equivalent. The MultiStage2.py
1075 script that generates these tables (the pcre_ucd.c file) now scans
1076 CaseFolding.txt instead of UnicodeData.txt for character case
1077 information.
1079 (b) The code for adding characters or ranges of characters to a character
1080 class has been abstracted into a generalized function that also handles
1081 case-independence. In UTF-mode with UCP support, this uses the new data
1082 to handle characters with more than one other case.
1084 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1085 whose other case is greater than 256 are now correctly matched
1086 caselessly. Previously, the high codepoint matched the low one, but not
1087 vice versa.
1089 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1090 of the new class addition function, using character lists defined as
1091 macros alongside the case definitions of 20 above.
1093 (e) Caseless back references now work with characters that have more than
1094 one other case.
1096 (f) General caseless matching of characters with more than one other case
1097 is supported.
1099 22. Unicode character properties were updated from Unicode 6.2.0
1101 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1103 24. Add support for 32-bit character strings, and UTF-32
1105 25. Major JIT compiler update (code refactoring and bugfixing).
1106 Experimental Sparc 32 support is added.
1108 26. Applied a modified version of Daniel Richard G's patch to create
1109 pcre.h.generic and config.h.generic by "make" instead of in the
1110 PrepareRelease script.
1112 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1113 pcre_compile.c when checking for a zero character.
1115 28. Introducing a native interface for JIT. Through this interface, the compiled
1116 machine code can be directly executed. The purpose of this interface is to
1117 provide fast pattern matching, so several sanity checks are not performed.
1118 However, feature tests are still performed. The new interface provides
1119 1.4x speedup compared to the old one.
1121 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1122 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1123 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1125 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1126 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1127 the "old" RFC 2279). Instead, it ended up passing a negative length to
1128 pcre_exec().
1130 31. Add support for GCC's visibility feature to hide internal functions.
1132 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1133 "unknown -C option" after outputting 0 or 1.
1135 33. There is now support for generating a code coverage report for the test
1136 suite in environments where gcc is the compiler and lcov is installed. This
1137 is mainly for the benefit of the developers.
1139 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1140 unaddressable using valgrind annotations, allowing valgrind to detect
1141 invalid memory accesses. This is mainly for the benefit of the developers.
1143 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1145 26. Give configure error if --enable-cpp but no C++ compiler found.
1148 Version 8.31 06-July-2012
1149 -------------------------
1151 1. Fixing a wrong JIT test case and some compiler warnings.
1153 2. Removed a bashism from the RunTest script.
1155 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1156 to unsigned type, result still unsigned" that was given by an MS compiler
1157 on encountering the code "-sizeof(xxx)".
1159 4. Partial matching support is added to the JIT compiler.
1161 5. Fixed several bugs concerned with partial matching of items that consist
1162 of more than one character:
1164 (a) /^(..)\1/ did not partially match "aba" because checking references was
1165 done on an "all or nothing" basis. This also applied to repeated
1166 references.
1168 (b) \R did not give a hard partial match if \r was found at the end of the
1169 subject.
1171 (c) \X did not give a hard partial match after matching one or more
1172 characters at the end of the subject.
1174 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1175 a partial match for the string "\r".
1177 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1178 a partial match for a CR character at the end of the subject string.
1180 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1181 running pcretest, the text "(JIT)" added to the output whenever JIT is
1182 actually used to run the match.
1184 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1185 or /S+[+] with a digit between 1 and 7.
1187 8. OP_NOT now supports any UTF character not just single-byte ones.
1189 9. (*MARK) control verb is now supported by the JIT compiler.
1191 10. The command "./RunTest list" lists the available tests without actually
1192 running any of them. (Because I keep forgetting what they all are.)
1196 12. Applied a (slightly modified) user-supplied patch that improves performance
1197 when the heap is used for recursion (compiled with --disable-stack-for-
1198 recursion). Instead of malloc and free for each heap frame each time a
1199 logical recursion happens, frames are retained on a chain and re-used where
1200 possible. This sometimes gives as much as 30% improvement.
1202 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1203 call.
1205 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1207 15. It is now possible to link pcretest with libedit as an alternative to
1208 libreadline.
1210 16. (*COMMIT) control verb is now supported by the JIT compiler.
1212 17. The Unicode data tables have been updated to Unicode 6.1.0.
1214 18. Added --file-list option to pcregrep.
1216 19. Added binary file support to pcregrep, including the -a, --binary-files,
1217 -I, and --text options.
1219 20. The madvise function is renamed for posix_madvise for QNX compatibility
1220 reasons. Fixed by Giuseppe D'Angelo.
1222 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1224 22. Changed the option for creating symbolic links for 16-bit man pages from
1225 -s to -sf so that re-installing does not cause issues.
1227 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1229 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1230 matches in certain environments (the workspace was not being correctly
1231 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1232 some of the workspace data at the beginning of a restart.
1234 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1235 was not doing so when it should - probably a typo introduced by SVN 528
1236 (change 8.10/14).
1238 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1239 \w+ when the character tables indicated that \x{c4} was a word character.
1240 There were several related cases, all because the tests for doing a table
1241 lookup were testing for characters less than 127 instead of 255.
1243 27. If a pattern contains capturing parentheses that are not used in a match,
1244 their slots in the ovector are set to -1. For those that are higher than
1245 any matched groups, this happens at the end of processing. In the case when
1246 there were back references that the ovector was too small to contain
1247 (causing temporary malloc'd memory to be used during matching), and the
1248 highest capturing number was not used, memory off the end of the ovector
1249 was incorrectly being set to -1. (It was using the size of the temporary
1250 memory instead of the true size.)
1252 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1253 ovector size, it uses memory at the end of the block that it has got.
1255 29. Check for an overlong MARK name and give an error at compile time. The
1256 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1258 30. JIT compiler update.
1260 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1261 Rill for the patch.
1263 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1265 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1267 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1268 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1269 was enabled.
1271 35. Improve JIT code generation for greedy plus quantifier.
1273 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1274 1 to "aa" instead of to an empty string. The bug affected repeated groups
1275 that could potentially match an empty string.
1277 37. Optimizing single character iterators in JIT.
1279 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1280 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1281 codepoints that are too big for the mode are faulted, and in a UTF mode,
1282 disallowed codepoints are also faulted.
1284 39. If PCRE was compiled with UTF support, in three places in the DFA
1285 matcher there was code that should only have been obeyed in UTF mode, but
1286 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1287 processing when bytes with values greater than 127 were present. In 16-bit
1288 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1289 both cases the values are those that cannot be the first data item in a UTF
1290 character. The three items that might have provoked this were recursions,
1291 possessively repeated groups, and atomic groups.
1293 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1294 and pcregrep, because some OS require shared objects to be explicitly
1295 passed to ld, causing the link step to fail if they are not.
1297 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1298 mode, patterns that started with \h* or \R* might be incorrectly matched.
1301 Version 8.30 04-February-2012
1302 -----------------------------
1304 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1305 name is defined in ctype.h.
1307 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1308 only in quite long subpatterns.
1310 3. Removed the function pcre_info(), which has been obsolete and deprecated
1311 since it was replaced by pcre_fullinfo() in February 2000.
1313 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1314 match a (*MARK), and the match failed at the start of the subject, a
1315 reference to memory before the start of the subject could occur. This bug
1316 was introduced by fix 17 of release 8.21.
1318 5. A reference to an unset group with zero minimum repetition was giving
1319 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1320 /(another)?(\1?)test/ matched against "hello world test". This bug was
1321 introduced in release 8.13.
1323 6. Add support for 16-bit character strings (a large amount of work involving
1324 many changes and refactorings).
1326 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1327 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1328 from a file.
1330 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1331 rounding is not applied in this particular case).
1333 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1334 if they appear, or are escaped, in patterns.
1336 10. Get rid of a number of -Wunused-but-set-variable warnings.
1338 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1339 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1340 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1341 also returns the mark "x". This bug applied to capturing parentheses,
1342 non-capturing parentheses, and atomic parentheses. It also applied to some
1343 assertions.
1345 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1346 information out of configure.ac instead of relying on pcre.h.generic, which
1347 is not stored in the repository.
1349 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1350 -lreadline.
1352 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1354 15. Applied Graycode's patch to put the top-level frame on the stack rather
1355 than the heap when not using the stack for recursion. This gives a
1356 performance improvement in many cases when recursion is not deep.
1358 16. Experimental code added to "pcretest -C" to output the stack frame size.
1361 Version 8.21 12-Dec-2011
1362 ------------------------
1364 1. Updating the JIT compiler.
1366 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1367 are added as well.
1369 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1370 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1371 calling _pcre_jit_exec. Some extra comments are added.
1373 4. (*MARK) settings inside atomic groups that do not contain any capturing
1374 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1375 was introduced by change 18 for 8.20.
1377 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1378 ECMA-262 standard.
1380 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1381 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1382 This bug was probably introduced by change 9 of 8.13.
1384 7. While fixing 6 above, I noticed that a number of other items were being
1385 incorrectly rejected as "not fixed length". This arose partly because newer
1386 opcodes had not been added to the fixed-length checking code. I have (a)
1387 corrected the bug and added tests for these items, and (b) arranged for an
1388 error to occur if an unknown opcode is encountered while checking for fixed
1389 length instead of just assuming "not fixed length". The items that were
1390 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1391 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1392 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1394 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1395 being incorrectly compiled and would have given unpredicatble results.
1397 9. A possessively repeated subpattern with minimum repeat count greater than
1398 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1399 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1400 the first (A) could occur when it should not.
1402 10. Add a cast and remove a redundant test from the code.
1404 11. JIT should use pcre_malloc/pcre_free for allocation.
1406 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1407 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1408 is anything other than /usr, -L is still shown).
1410 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1412 14. Perl does not support \N without a following name in a [] class; PCRE now
1413 also gives an error.
1415 15. If a forward reference was repeated with an upper limit of around 2000,
1416 it caused the error "internal error: overran compiling workspace". The
1417 maximum number of forward references (including repeats) was limited by the
1418 internal workspace, and dependent on the LINK_SIZE. The code has been
1419 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1420 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1421 of around 200,000 forward references. While doing this, I also speeded up
1422 the filling in of repeated forward references.
1424 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1425 incorrectly expecting the subject to contain another "a" after the start.
1427 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1428 in the match, the SKIP should be ignored. This was not happening; instead
1429 the SKIP was being treated as NOMATCH. For patterns such as
1430 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1431 tested.
1433 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1434 now much more compatible with Perl, in particular in cases where the result
1435 is a non-match for a non-anchored pattern. For example, if
1436 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1437 "m", where previously it did not return a name. A side effect of this
1438 change is that for partial matches, the last encountered mark name is
1439 returned, as for non matches. A number of tests that were previously not
1440 Perl-compatible have been moved into the Perl-compatible test files. The
1441 refactoring has had the pleasing side effect of removing one argument from
1442 the match() function, thus reducing its stack requirements.
1444 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1445 subsequent uses of /S (without +) incorrectly behaved like /S+.
1447 21. Retrieve executable code size support for the JIT compiler and fixing
1448 some warnings.
1450 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1451 not work when the shorter character appeared right at the end of the
1452 subject string.
1454 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1455 systems.
1457 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1458 output it when the /M option is used in pcretest.
1460 25. The CheckMan script was not being included in the distribution. Also, added
1461 an explicit "perl" to run Perl scripts from the PrepareRelease script
1462 because this is reportedly needed in Windows.
1464 26. If study data was being save in a file and studying had not found a set of
1465 "starts with" bytes for the pattern, the data written to the file (though
1466 never used) was taken from uninitialized memory and so caused valgrind to
1467 complain.
1469 27. Updated RunTest.bat as provided by Sheri Pierce.
1471 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1473 29. Computation of memory usage for the table of capturing group names was
1474 giving an unnecessarily large value.
1477 Version 8.20 21-Oct-2011
1478 ------------------------
1480 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1481 a POSIX class. After further experiments with Perl, which convinced me that
1482 Perl has bugs and confusions, a closing square bracket is no longer allowed
1483 in a POSIX name. This bug also affected patterns with classes that started
1484 with full stops.
1486 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1487 captured substring, but while checking the failing first alternative,
1488 substring 1 is temporarily captured. If the output vector supplied to
1489 pcre_exec() was not big enough for this capture, the yield of the function
1490 was still zero ("insufficient space for captured substrings"). This cannot
1491 be totally fixed without adding another stack variable, which seems a lot
1492 of expense for a edge case. However, I have improved the situation in cases
1493 such as /(a)(b)x|abc/ matched against "abc", where the return code
1494 indicates that fewer than the maximum number of slots in the ovector have
1495 been set.
1497 3. Related to (2) above: when there are more back references in a pattern than
1498 slots in the output vector, pcre_exec() uses temporary memory during
1499 matching, and copies in the captures as far as possible afterwards. It was
1500 using the entire output vector, but this conflicts with the specification
1501 that only 2/3 is used for passing back captured substrings. Now it uses
1502 only the first 2/3, for compatibility. This is, of course, another edge
1503 case.
1505 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1506 main code base, and can be used by building with --enable-jit. When this is
1507 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1508 runtime --no-jit option is given.
1510 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1511 ovector, the return from the function was zero, implying that there were
1512 other matches that did not fit. The correct "exactly full" value is now
1513 returned.
1515 6. If a subpattern that was called recursively or as a subroutine contained
1516 (*PRUNE) or any other control that caused it to give a non-standard return,
1517 invalid errors such as "Error -26 (nested recursion at the same subject
1518 position)" or even infinite loops could occur.
1520 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1521 computing the minimum length on reaching *ACCEPT, and so ended up with the
1522 wrong value of 1 rather than 0. Further investigation indicates that
1523 computing a minimum subject length in the presence of *ACCEPT is difficult
1524 (think back references, subroutine calls), and so I have changed the code
1525 so that no minimum is registered for a pattern that contains *ACCEPT.
1527 8. If (*THEN) was present in the first (true) branch of a conditional group,
1528 it was not handled as intended. [But see 16 below.]
1530 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1531 Sheri Pierce.
1533 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1534 the first byte in a match must be "a".
1536 11. Change 17 for 8.13 increased the recursion depth for patterns like
1537 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1538 pattern contains any instances of (*THEN). If it does not, the old
1539 optimizations are restored. It would be nice to do this on a per-group
1540 basis, but at the moment that is not feasible.
1542 12. In some environments, the output of pcretest -C is CRLF terminated. This
1543 broke RunTest's code that checks for the link size. A single white space
1544 character after the value is now allowed for.
1546 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1547 For "fr", it uses the Windows-specific input and output files.
1549 14. If (*THEN) appeared in a group that was called recursively or as a
1550 subroutine, it did not work as intended. [But see next item.]
1552 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1553 pattern fragments (but not containing any | characters). If A and B are
1554 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1555 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1556 D. In other words, Perl considers parentheses that do not contain any |
1557 characters to be part of a surrounding alternative, whereas PCRE was
1558 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1559 differently. PCRE now behaves in the same way as Perl, except in the case
1560 of subroutine/recursion calls such as (?1) which have in any case always
1561 been different (but PCRE had them first :-).
1563 16. Related to 15 above: Perl does not treat the | in a conditional group as
1564 creating alternatives. Such a group is treated in the same way as an
1565 ordinary group without any | characters when processing (*THEN). PCRE has
1566 been changed to match Perl's behaviour.
1568 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1569 RunGrepTest script failed.
1571 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1572 inevitable for groups that contain captures, but it can lead to a lot of
1573 stack use in large patterns. The old behaviour has been restored for atomic
1574 groups that do not contain any capturing parentheses.
1576 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1577 suppress the check for a minimum subject length at run time. (If it was
1578 given to pcre_exec() or pcre_dfa_exec() it did work.)
1580 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1581 fail to work when decoding hex characters in data strings in EBCDIC
1582 environments.
1584 21. It appears that in at least one Mac OS environment, the isxdigit() function
1585 is implemented as a macro that evaluates to its argument more than once,
1586 contravening the C 90 Standard (I haven't checked a later standard). There
1587 was an instance in pcretest which caused it to go wrong when processing
1588 \x{...} escapes in subject strings. The has been rewritten to avoid using
1589 things like p++ in the argument of isxdigit().
1592 Version 8.13 16-Aug-2011
1593 ------------------------
1595 1. The Unicode data tables have been updated to Unicode 6.0.0.
1597 2. Two minor typos in pcre_internal.h have been fixed.
1599 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1600 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1601 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1603 4. There were a number of related bugs in the code for matching backrefences
1604 caselessly in UTF-8 mode when codes for the characters concerned were
1605 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1606 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1607 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1608 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1609 2-byte code at the end of the subject (it thought there wasn't enough data
1610 left).
1612 5. Comprehensive information about what went wrong is now returned by
1613 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1614 as the output vector has at least 2 elements. The offset of the start of
1615 the failing character and a reason code are placed in the vector.
1617 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1618 now returned is for the first byte of the failing character, instead of the
1619 last byte inspected. This is an incompatible change, but I hope it is small
1620 enough not to be a problem. It makes the returned offset consistent with
1621 pcre_exec() and pcre_dfa_exec().
1623 7. pcretest now gives a text phrase as well as the error number when
1624 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1625 failure, the offset and reason code are output.
1627 8. When \R was used with a maximizing quantifier it failed to skip backwards
1628 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1629 back over a single character (\n). This seems wrong (because it treated the
1630 two characters as a single entity when going forwards), conflicts with the
1631 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1632 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1633 has been changed.
1635 9. Some internal refactoring has changed the processing so that the handling
1636 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1637 time (the PCRE_DOTALL option was changed this way some time ago: version
1638 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1639 which was always a bit of a fudge. It also means that there is one less
1640 argument for the match() function, which reduces its stack requirements
1641 slightly. This change also fixes an incompatibility with Perl: the pattern
1642 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1644 10. More internal refactoring has drastically reduced the number of recursive
1645 calls to match() for possessively repeated groups such as (abc)++ when
1646 using pcre_exec().
1648 11. While implementing 10, a number of bugs in the handling of groups were
1649 discovered and fixed:
1651 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1652 (a|)*(?1) gave a compile-time internal error.
1653 ((a|)+)+ did not notice that the outer group could match an empty string.
1654 (^a|^)+ was not marked as anchored.
1655 (.*a|.*)+ was not marked as matching at start or after a newline.
1657 12. Yet more internal refactoring has removed another argument from the match()
1658 function. Special calls to this function are now indicated by setting a
1659 value in a variable in the "match data" data block.
1661 13. Be more explicit in pcre_study() instead of relying on "default" for
1662 opcodes that mean there is no starting character; this means that when new
1663 ones are added and accidentally left out of pcre_study(), testing should
1664 pick them up.
1666 14. The -s option of pcretest has been documented for ages as being an old
1667 synonym of -m (show memory usage). I have changed it to mean "force study
1668 for every regex", that is, assume /S for every regex. This is similar to -i
1669 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1670 using it. It makes it easier to run collections of tests with and without
1671 study enabled, and thereby test pcre_study() more easily. All the standard
1672 tests are now run with and without -s (but some patterns can be marked as
1673 "never study" - see 20 below).
1675 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1676 restoration of the capturing data to the outer values was not happening
1677 correctly.
1679 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1680 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1681 pattern had matched an empty string, and so incorrectly returned a no
1682 match.
1684 17. There was optimizing code for the last branch of non-capturing parentheses,
1685 and also for the obeyed branch of a conditional subexpression, which used
1686 tail recursion to cut down on stack usage. Unfortunately, now that there is
1687 the possibility of (*THEN) occurring in these branches, tail recursion is
1688 no longer possible because the return has to be checked for (*THEN). These
1689 two optimizations have therefore been removed. [But see 8.20/11 above.]
1691 18. If a pattern containing \R was studied, it was assumed that \R always
1692 matched two bytes, thus causing the minimum subject length to be
1693 incorrectly computed because \R can also match just one byte.
1695 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1696 was incorrectly computed.
1698 20. If /S is present twice on a test pattern in pcretest input, it now
1699 *disables* studying, thereby overriding the use of -s on the command line
1700 (see 14 above). This is necessary for one or two tests to keep the output
1701 identical in both cases.
1703 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1704 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1706 22. When an atomic group that contained a capturing parenthesis was
1707 successfully matched, but the branch in which it appeared failed, the
1708 capturing was not being forgotten if a higher numbered group was later
1709 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1710 group 1 to "a", when in fact it should be unset. This applied to multi-
1711 branched capturing and non-capturing groups, repeated or not, and also to
1712 positive assertions (capturing in negative assertions does not happen
1713 in PCRE) and also to nested atomic groups.
1715 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1716 subject after a captured substring, to make it easier to tell which of a
1717 number of identical substrings has been captured.
1719 24. The way atomic groups are processed by pcre_exec() has been changed so that
1720 if they are repeated, backtracking one repetition now resets captured
1721 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1722 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1723 "aaa". Previously, it would have been "a". As part of this code
1724 refactoring, the way recursive calls are handled has also been changed.
1726 25. If an assertion condition captured any substrings, they were not passed
1727 back unless some other capturing happened later. For example, if
1728 (?(?=(a))a) was matched against "a", no capturing was returned.
1730 26. When studying a pattern that contained subroutine calls or assertions,
1731 the code for finding the minimum length of a possible match was handling
1732 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1733 group 1 called group 2 while simultaneously a separate group 2 called group
1734 1). A stack overflow occurred in this case. I have fixed this by limiting
1735 the recursion depth to 10.
1737 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1738 Fortmann. This supports explicit test numbers on the command line, and has
1739 argument validation and error reporting.
1741 28. An instance of \X with an unlimited repeat could fail if at any point the
1742 first character it looked at was a mark character.
1744 29. Some minor code refactoring concerning Unicode properties and scripts
1745 should reduce the stack requirement of match() slightly.
1747 30. Added the '=' option to pcretest to check the setting of unused capturing
1748 slots at the end of the pattern, which are documented as being -1, but are
1749 not included in the return count.
1751 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1752 compiled something random. Now it gives a compile-time error (as does
1753 Perl).
1755 32. A *MARK encountered during the processing of a positive assertion is now
1756 recorded and passed back (compatible with Perl).
1758 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1759 had alternative anchored branches, the search for a second match in a line
1760 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1761 matched the line "0102" twice. The same bug affected patterns that started
1762 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1763 twice.
1765 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1766 does, and because of capturing effects, quantifying parenthesized
1767 assertions may at times be useful. Quantifiers are now allowed for
1768 parenthesized assertions.
1770 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1772 36. \g was being checked for fancy things in a character class, when it should
1773 just be a literal "g".
1775 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1776 appearance of a nested POSIX class supersedes an apparent external class.
1777 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1778 unescaped square brackets may also appear as part of class names. For
1779 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1780 more like Perl. (But see 8.20/1 above.)
1782 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1783 was because it thought it was \N{name}, which is not supported).
1785 39. Add minix to OS list not supporting the -S option in pcretest.
1787 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1788 cannot analyze patterns in sufficient detail to catch mutual recursions
1789 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1790 subgroup is called recursively as a subpattern for a second time at the
1791 same position in the subject string. In previous releases this might have
1792 been caught by the recursion limit, or it might have run out of stack.
1794 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1795 happen only once. PCRE was, however incorrectly giving a compile time error
1796 "recursive call could loop indefinitely" because it cannot analyze the
1797 pattern in sufficient detail. The compile time test no longer happens when
1798 PCRE is compiling a conditional subpattern, but actual runaway loops are
1799 now caught at runtime (see 40 above).
1801 42. It seems that Perl allows any characters other than a closing parenthesis
1802 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1803 has been changed to be the same.
1805 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1806 as not to get warnings when autogen.sh is called. Also changed
1807 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1809 44. To help people who use pcregrep to scan files containing exceedingly long
1810 lines, the following changes have been made:
1812 (a) The default value of the buffer size parameter has been increased from
1813 8K to 20K. (The actual buffer used is three times this size.)
1815 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1816 PCRE is built.
1818 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1819 to be set at run time.
1821 (d) Numerical values in pcregrep options can be followed by K or M, for
1822 example --buffer-size=50K.
1824 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1825 given and the return code is set to 2.
1827 45. Add a pointer to the latest mark to the callout data block.
1829 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1830 partial match of an empty string instead of no match. This was specific to
1831 the use of ".".
1833 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1834 complete match instead of a partial match. This bug was dependent on both
1835 the PCRE_UTF8 and PCRE_DOTALL options being set.
1837 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1838 starting byte set, because \b was not being ignored.
1841 Version 8.12 15-Jan-2011
1842 ------------------------
1844 1. Fixed some typos in the markup of the man pages, and wrote a script that
1845 checks for such things as part of the documentation building process.
1847 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1848 --match-limit and --recursion-limit options (added for 8.11). In
1849 particular, this made one of the standard tests fail. (The integer value
1850 went into the wrong half of a long int.)
1852 3. If the --colour option was given to pcregrep with -v (invert match), it
1853 did strange things, either producing crazy output, or crashing. It should,
1854 of course, ignore a request for colour when reporting lines that do not
1855 match.
1857 4. Another pcregrep bug caused similar problems if --colour was specified with
1858 -M (multiline) and the pattern match finished with a line ending.
1860 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1861 matched in multiline mode, the following line was shown as part of the
1862 match. This seems wrong, so I have changed it.
1864 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1865 the check for further matches in the same line (so they could be coloured)
1866 to overrun the end of the current line. If another match was found, it was
1867 incorrectly shown (and then shown again when found in the next line).
1869 7. If pcregrep was compiled under Windows, there was a reference to the
1870 function pcregrep_exit() before it was defined. I am assuming this was
1871 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1872 reported by a user. I've moved the definition above the reference.
1875 Version 8.11 10-Dec-2010
1876 ------------------------
1878 1. (*THEN) was not working properly if there were untried alternatives prior
1879 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1880 backtracked to try for "b" instead of moving to the next alternative branch
1881 at the same level (in this case, to look for "c"). The Perl documentation
1882 is clear that when (*THEN) is backtracked onto, it goes to the "next
1883 alternative in the innermost enclosing group".
1885 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1886 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1887 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1888 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1889 (*THEN).
1891 3. If \s appeared in a character class, it removed the VT character from
1892 the class, even if it had been included by some previous item, for example
1893 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1894 of \s, but is part of the POSIX "space" class.)
1896 4. A partial match never returns an empty string (because you can always
1897 match an empty string at the end of the subject); however the checking for
1898 an empty string was starting at the "start of match" point. This has been
1899 changed to the "earliest inspected character" point, because the returned
1900 data for a partial match starts at this character. This means that, for
1901 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1902 (previously it gave "no match").
1904 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1905 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1906 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1907 has an implication that the given string is incomplete (because a partial
1908 match is preferred over a full match). For this reason, these items now
1909 give a partial match in this situation. [Aside: previously, the one case
1910 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1911 match rather than a full match, which was wrong by the old rules, but is
1912 now correct.]
1914 6. There was a bug in the handling of #-introduced comments, recognized when
1915 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1916 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1917 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1918 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1919 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1920 places in pcre_compile().
1922 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1923 comments when looking ahead for named forward references to subpatterns,
1924 the only newline sequence it recognized was NL. It now handles newlines
1925 according to the set newline convention.
1927 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1928 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1929 cater for a lack of strerror(). These oversights have been fixed.
1931 9. Added --match-limit and --recursion-limit to pcregrep.
1933 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1935 11. When the -o option was used, pcregrep was setting a return code of 1, even
1936 when matches were found, and --line-buffered was not being honoured.
1938 12. Added an optional parentheses number to the -o and --only-matching options
1939 of pcregrep.
1941 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1942 can match an empty string. The code to do it in pcretest and pcredemo
1943 needed fixing:
1945 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1946 only one byte after an empty string match just before CRLF (this case
1947 just got forgotten; "any" and "anycrlf" were OK).
1949 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1950 mode when an empty string match preceded an ASCII character followed by
1951 a non-ASCII character. (The code for advancing by one character rather
1952 than one byte was nonsense.)
1954 (c) The pcredemo.c sample program did not have any code at all to handle
1955 the cases when CRLF is a valid newline sequence.
1957 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1958 as a starting offset was within the subject string. There is now a new
1959 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1960 negative or greater than the length of the string. In order to test this,
1961 pcretest is extended to allow the setting of negative starting offsets.
1963 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1964 starting offset points to the beginning of a UTF-8 character was
1965 unnecessarily clumsy. I tidied it up.
1967 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1968 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1970 17. Nobody had reported that the --include_dir option, which was added in
1971 release 7.7 should have been called --include-dir (hyphen, not underscore)
1972 for compatibility with GNU grep. I have changed it to --include-dir, but
1973 left --include_dir as an undocumented synonym, and the same for
1974 --exclude-dir, though that is not available in GNU grep, at least as of
1975 release 2.5.4.
1977 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1978 characters from a string of bytes) have been redefined so as not to use
1979 loops, in order to improve performance in some environments. At the same
1980 time, I abstracted some of the common code into auxiliary macros to save
1981 repetition (this should not affect the compiled code).
1983 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1984 compile-time error is now given if \c is not followed by an ASCII
1985 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1986 different, and any byte value is allowed.)
1988 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1989 START_OPTIMIZE option, which is now allowed at compile time - but just
1990 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1991 to pcregrep and other applications that have no direct access to PCRE
1992 options. The new /Y option in pcretest sets this option when calling
1993 pcre_compile().
1995 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1996 back references. Groups containing recursive back references were forced to
1997 be atomic by that change, but in the case of named groups, the amount of
1998 memory required was incorrectly computed, leading to "Failed: internal
1999 error: code overflow". This has been fixed.
2001 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2002 pcretest.c, to avoid build problems in some Borland environments.
2005 Version 8.10 25-Jun-2010
2006 ------------------------
2008 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2009 THEN.
2011 2. (*ACCEPT) was not working when inside an atomic group.
2013 3. Inside a character class, \B is treated as a literal by default, but
2014 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2015 causes the error). The code is unchanged, but I tidied the documentation.
2017 4. Inside a character class, PCRE always treated \R and \X as literals,
2018 whereas Perl faults them if its -w option is set. I have changed PCRE so
2019 that it faults them when PCRE_EXTRA is set.
2021 5. Added support for \N, which always matches any character other than
2022 newline. (It is the same as "." when PCRE_DOTALL is not set.)
2024 6. When compiling pcregrep with newer versions of gcc which may have
2025 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2026 declared with attribute warn_unused_result" were given. Just casting the
2027 result to (void) does not stop the warnings; a more elaborate fudge is
2028 needed. I've used a macro to implement this.
2030 7. Minor change to pcretest.c to avoid a compiler warning.
2032 8. Added four artifical Unicode properties to help with an option to make
2033 \s etc use properties (see next item). The new properties are: Xan
2034 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2036 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2037 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2038 this option. Modified pcretest to add /W to test this facility. Added
2039 REG_UCP to make it available via the POSIX interface.
2041 10. Added --line-buffered to pcregrep.
2043 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2044 studied, and the match started with a letter with a code point greater than
2045 127 whose first byte was different to the first byte of the other case of
2046 the letter, the other case of this starting letter was not recognized
2047 (#976).
2049 12. If a pattern that was studied started with a repeated Unicode property
2050 test, for example, \p{Nd}+, there was the theoretical possibility of
2051 setting up an incorrect bitmap of starting bytes, but fortunately it could
2052 not have actually happened in practice until change 8 above was made (it
2053 added property types that matched character-matching opcodes).
2055 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2056 possible starting bytes for non-anchored patterns.
2058 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2059 \R, and also a number of cases that involve Unicode properties, both
2060 explicit and implicit when PCRE_UCP is set.
2062 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2063 input, it could crash or give wrong results if characters with values
2064 greater than 0xc0 were present in the subject string. (Detail: it assumed
2065 UTF-8 input when processing these items.)
2067 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2068 size_t is 64-bit (#991).
2070 17. Added a check for running out of memory when PCRE is compiled with
2071 --disable-stack-for-recursion (#990).
2073 18. If the last data line in a file for pcretest does not have a newline on
2074 the end, a newline was missing in the output.
2076 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2077 less than 128) in its various bitmaps. However, there is a facility for
2078 generating tables according to the current locale when PCRE is compiled. It
2079 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2080 characters, are recognized by isspace() and therefore were getting set in
2081 these tables, and indeed these tables seem to approximate to ISO 8859. This
2082 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2083 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2084 which of course cannot start UTF-8 characters. I have changed the code so
2085 that only real ASCII characters (less than 128) and the correct starting
2086 bytes for UTF-8 encodings are set for characters greater than 127 when in
2087 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2088 altogether.)
2090 20. Added the /T option to pcretest so as to be able to run tests with non-
2091 standard character tables, thus making it possible to include the tests
2092 used for 19 above in the standard set of tests.
2094 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2095 reference to a subpattern the other side of a comment that contains an
2096 opening parenthesis caused either an internal compiling error, or a
2097 reference to the wrong subpattern.
2100 Version 8.02 19-Mar-2010
2101 ------------------------
2103 1. The Unicode data tables have been updated to Unicode 5.2.0.
2105 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2106 configured.
2108 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2109 original author of that file, following a query about its status.
2111 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2112 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2114 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2115 quantifier applied to a forward-referencing subroutine call, could compile
2116 incorrect code or give the error "internal error: previously-checked
2117 referenced subpattern not found".
2119 6. Both MS Visual Studio and Symbian OS have problems with initializing
2120 variables to point to external functions. For these systems, therefore,
2121 pcre_malloc etc. are now initialized to local functions that call the
2122 relevant global functions.
2124 7. There were two entries missing in the vectors called coptable and poptable
2125 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2126 I've fixed the data, and added a kludgy way of testing at compile time that
2127 the lengths are correct (equal to the number of opcodes).
2129 8. Following on from 7, I added a similar kludge to check the length of the
2130 eint vector in pcreposix.c.
2132 9. Error texts for pcre_compile() are held as one long string to avoid too
2133 much relocation at load time. To find a text, the string is searched,
2134 counting zeros. There was no check for running off the end of the string,
2135 which could happen if a new error number was added without updating the
2136 string.
2138 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2140 11. \K was not working if it appeared in an atomic group or in a group that
2141 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2142 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2143 the assertion is positive, but not if it is negative.
2145 12. Change 11 fortuitously reduced the size of the stack frame used in the
2146 "match()" function of pcre_exec.c by one pointer. Forthcoming
2147 implementation of support for (*MARK) will need an extra pointer on the
2148 stack; I have reserved it now, so that the stack frame size does not
2149 decrease.
2151 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2152 item in branch that calls a recursion is a subroutine call - as in the
2153 second branch in the above example - was incorrectly given the compile-
2154 time error "recursive call could loop indefinitely" because pcre_compile()
2155 was not correctly checking the subroutine for matching a non-empty string.
2157 14. The checks for overrunning compiling workspace could trigger after an
2158 overrun had occurred. This is a "should never occur" error, but it can be
2159 triggered by pathological patterns such as hundreds of nested parentheses.
2160 The checks now trigger 100 bytes before the end of the workspace.
2162 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2165 Version 8.01 19-Jan-2010
2166 ------------------------
2168 1. If a pattern contained a conditional subpattern with only one branch (in
2169 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2170 computed the wrong minimum data length (which is of course zero for such
2171 subpatterns). This could cause incorrect "no match" results.
2173 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2174 the pattern is reset in the first branch, pcre_compile() failed with
2175 "internal error: code overflow at offset...". This happened only when
2176 the reset was to the original external option setting. (An optimization
2177 abstracts leading options settings into an external setting, which was the
2178 cause of this.)
2180 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2181 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2182 assertion pattern did not match (meaning that the assertion was true), it
2183 was incorrectly treated as false if the SKIP had been reached during the
2184 matching. This also applied to assertions used as conditions.
2186 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2187 assertion subpattern, including such a pattern used as a condition,
2188 unpredictable results occurred, instead of the error return
2191 5. The C++ GlobalReplace function was not working like Perl for the special
2192 situation when an empty string is matched. It now does the fancy magic
2193 stuff that is necessary.
2195 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2196 removed. (These were left over from very, very early versions of PCRE.)
2198 7. Some cosmetic changes to the code to make life easier when compiling it
2199 as part of something else:
2201 (a) Change DEBUG to PCRE_DEBUG.
2203 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2204 called "current" as "current_branch", to prevent a collision with the
2205 Linux macro when compiled as a kernel module.
2207 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2208 prevent a collision with the Linux macro when compiled as a kernel
2209 module.
2211 8. In pcre_compile() there are some checks for integer overflows that used to
2212 cast potentially large values to (double). This has been changed to that
2213 when building, a check for int64_t is made, and if it is found, it is used
2214 instead, thus avoiding the use of floating point arithmetic. (There is no
2215 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2216 double.
2218 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2219 2005 (difference between two addresses compared to an unsigned value).
2221 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2222 custom one, because of the following reported problem in Windows:
2224 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2225 under Win32.
2226 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2227 therefore missing the function definition.
2228 - The compiler thus generates a "C" signature for the test function.
2229 - The linker fails to find the "C" function.
2230 - PCRE fails to configure if asked to do so against libbz2.
2232 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2233 messages were output:
2235 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2236 rerunning libtoolize, to keep the correct libtool macros in-tree.
2237 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2239 I have done both of these things.
2241 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2242 most of the time, it *can* run out if it is given a pattern that contains a
2243 runaway infinite recursion. I updated the discussion in the pcrestack man
2244 page.
2246 13. Now that we have gone to the x.xx style of version numbers, the minor
2247 version may start with zero. Using 08 or 09 is a bad idea because users
2248 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2249 interpreted as invalid octal numbers. I've updated the previous comment in
2250 configure.ac, and also added a check that gives an error if 08 or 09 are
2251 used.
2253 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2254 causing partial matching to fail when the end of the subject matched \W
2255 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2257 15. There were some discrepancies between the declarations in pcre_internal.h
2258 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2259 their definitions. The declarations used "const uschar *" and the
2260 definitions used USPTR. Even though USPTR is normally defined as "const
2261 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2262 reported that: "This difference in casting confuses some C++ compilers, for
2263 example, SunCC recognizes above declarations as different functions and
2264 generates broken code for hbpcre." I have changed the declarations to use
2265 USPTR.
2267 16. GNU libtool is named differently on some systems. The autogen.sh script now
2268 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2269 (FreeBSD).
2271 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2272 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2273 comment: "Figure out how to create a longlong from a string: strtoll and
2274 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2275 instance, but it only takes 2 args instead of 3!"
2277 18. A subtle bug concerned with back references has been fixed by a change of
2278 specification, with a corresponding code fix. A pattern such as
2279 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2280 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2281 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2282 same bug. Such groups have to be quantified to be useful, or contained
2283 inside another quantified group. (If there's no repetition, the reference
2284 can never match.) The problem arises because, having left the group and
2285 moved on to the rest of the pattern, a later failure that backtracks into
2286 the group uses the captured value from the final iteration of the group
2287 rather than the correct earlier one. I have fixed this in PCRE by forcing
2288 any group that contains a reference to itself to be an atomic group; that
2289 is, there cannot be any backtracking into it once it has completed. This is
2290 similar to recursive and subroutine calls.
2293 Version 8.00 19-Oct-09
2294 ----------------------
2296 1. The table for translating pcre_compile() error codes into POSIX error codes
2297 was out-of-date, and there was no check on the pcre_compile() error code
2298 being within the table. This could lead to an OK return being given in
2299 error.
2301 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2302 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2303 in a Windows environment.
2305 3. The pcregrep --count option prints the count for each file even when it is
2306 zero, as does GNU grep. However, pcregrep was also printing all files when
2307 --files-with-matches was added. Now, when both options are given, it prints
2308 counts only for those files that have at least one match. (GNU grep just
2309 prints the file name in this circumstance, but including the count seems
2310 more useful - otherwise, why use --count?) Also ensured that the
2311 combination -clh just lists non-zero counts, with no names.
2313 4. The long form of the pcregrep -F option was incorrectly implemented as
2314 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2315 but it seems right to fix it, and I didn't think it was worth preserving
2316 the old behaviour.
2318 5. The command line items --regex=pattern and --regexp=pattern were not
2319 recognized by pcregrep, which required --regex pattern or --regexp pattern
2320 (with a space rather than an '='). The man page documented the '=' forms,
2321 which are compatible with GNU grep; these now work.
2323 6. No libpcreposix.pc file was created for pkg-config; there was just
2324 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2326 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2327 when UCP support is not needed, by modifying the Python script that
2328 generates it from Unicode data files. This should not matter if the module
2329 is correctly used as a library, but I received one complaint about 50K of
2330 unwanted data. My guess is that the person linked everything into his
2331 program rather than using a library. Anyway, it does no harm.
2333 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2334 was a minimum greater than 1 for a wide character in a possessive
2335 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2336 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2337 character. Chaos in the form of incorrect output or a compiling loop could
2338 result.
2340 9. The restrictions on what a pattern can contain when partial matching is
2341 requested for pcre_exec() have been removed. All patterns can now be
2342 partially matched by this function. In addition, if there are at least two
2343 slots in the offset vector, the offset of the earliest inspected character
2344 for the match and the offset of the end of the subject are set in them when
2345 PCRE_ERROR_PARTIAL is returned.
2347 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2348 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2349 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2350 and may be more useful for multi-segment matching.
2352 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2353 used to be given if ever the end of the subject was reached; now it is
2354 given only if matching could not proceed because another character was
2355 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2356 string "Z", which now yields "no match" instead of "partial match". In the
2357 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2358 final character ended with (*FAIL).
2360 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2361 if the pattern had a "must contain" character that was already found in the
2362 earlier partial match, unless partial matching was again requested. For
2363 example, with the pattern /dog.(body)?/, the "must contain" character is
2364 "g". If the first part-match was for the string "dog", restarting with
2365 "sbody" failed. This bug has been fixed.
2367 13. The string returned by pcre_dfa_exec() after a partial match has been
2368 changed so that it starts at the first inspected character rather than the
2369 first character of the match. This makes a difference only if the pattern
2370 starts with a lookbehind assertion or \b or \B (\K is not supported by
2371 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2372 matching functions compatible, and I think it's the right thing to do.
2374 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2375 so that the demonstration program is easily available in environments where
2376 PCRE has not been installed from source.
2378 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2379 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2380 library.
2382 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2383 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2384 is not the first non-POSIX option to be added. Clearly some people find
2385 these options useful.
2387 17. If a caller to the POSIX matching function regexec() passes a non-zero
2388 value for nmatch with a NULL value for pmatch, the value of
2389 nmatch is forced to zero.
2391 18. RunGrepTest did not have a test for the availability of the -u option of
2392 the diff command, as RunTest does. It now checks in the same way as
2393 RunTest, and also checks for the -b option.
2395 19. If an odd number of negated classes containing just a single character
2396 interposed, within parentheses, between a forward reference to a named
2397 subpattern and the definition of the subpattern, compilation crashed with
2398 an internal error, complaining that it could not find the referenced
2399 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2400 [The bug was that it was starting one character too far in when skipping
2401 over the character class, thus treating the ] as data rather than
2402 terminating the class. This meant it could skip too much.]
2404 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2405 /g option in pcretest when the pattern contains \K, which makes it possible
2406 to have an empty string match not at the start, even when the pattern is
2407 anchored. Updated pcretest and pcredemo to use this option.
2409 21. If the maximum number of capturing subpatterns in a recursion was greater
2410 than the maximum at the outer level, the higher number was returned, but
2411 with unset values at the outer level. The correct (outer level) value is
2412 now given.
2414 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2415 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2416 make it do so. The string so far is captured, making this feature
2417 compatible with Perl.
2419 23. The tests have been re-organized, adding tests 11 and 12, to make it
2420 possible to check the Perl 5.10 features against Perl 5.10.
2422 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2423 pattern matches a fixed length string. PCRE did not allow this; now it
2424 does. Neither allows recursion.
2426 25. I finally figured out how to implement a request to provide the minimum
2427 length of subject string that was needed in order to match a given pattern.
2428 (It was back references and recursion that I had previously got hung up
2429 on.) This code has now been added to pcre_study(); it finds a lower bound
2430 to the length of subject needed. It is not necessarily the greatest lower
2431 bound, but using it to avoid searching strings that are too short does give
2432 some useful speed-ups. The value is available to calling programs via
2433 pcre_fullinfo().
2435 26. While implementing 25, I discovered to my embarrassment that pcretest had
2436 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2437 study optimizations had never been tested with that matching function.
2438 Oops. What is worse, even when it was passed study data, there was a bug in
2439 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2440 were also very few tests of studied patterns with pcre_dfa_exec().
2442 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2443 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2444 on the other side of the coin, they are no longer allowed to have different
2445 names, because these cannot be distinguished in PCRE, and this has caused
2446 confusion. (This is a difference from Perl.)
2448 28. When duplicate subpattern names are present (necessarily with different
2449 numbers, as required by 27 above), and a test is made by name in a
2450 conditional pattern, either for a subpattern having been matched, or for
2451 recursion in such a pattern, all the associated numbered subpatterns are
2452 tested, and the overall condition is true if the condition is true for any
2453 one of them. This is the way Perl works, and is also more like the way
2454 testing by number works.
2457 Version 7.9 11-Apr-09
2458 ---------------------
2460 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2461 (pcretest), all targets were linked against these libraries. This included
2462 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2463 libraries. This caused unwanted dependencies to be created. This problem
2464 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2465 pcretest is linked with readline.
2467 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2468 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2469 moved outside it again, because FALSE and TRUE are already defined in AIX,
2470 but BOOL is not.
2472 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2473 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2475 4. The pcregrep documentation said spaces were inserted as well as colons (or
2476 hyphens) following file names and line numbers when outputting matching
2477 lines. This is not true; no spaces are inserted. I have also clarified the
2478 wording for the --colour (or --color) option.
2480 5. In pcregrep, when --colour was used with -o, the list of matching strings
2481 was not coloured; this is different to GNU grep, so I have changed it to be
2482 the same.
2484 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2485 each matching line was coloured. Now it goes on to look for further matches
2486 of any of the test patterns, which is the same behaviour as GNU grep.
2488 7. A pattern that could match an empty string could cause pcregrep to loop; it
2489 doesn't make sense to accept an empty string match in pcregrep, so I have
2490 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2491 seems to be how GNU grep behaves. [But see later change 40 for release
2492 8.33.]
2494 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2495 start or after a newline", because the conditional assertion was not being
2496 correctly handled. The rule now is that both the assertion and what follows
2497 in the first alternative must satisfy the test.
2499 9. If auto-callout was enabled in a pattern with a conditional group whose
2500 condition was an assertion, PCRE could crash during matching, both with
2501 pcre_exec() and pcre_dfa_exec().
2503 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2504 used for matching.
2506 11. Unicode property support in character classes was not working for
2507 characters (bytes) greater than 127 when not in UTF-8 mode.
2509 12. Added the -M command line option to pcretest.
2511 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2513 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2515 16. Added comments and documentation about mis-use of no_arg in the C++
2516 wrapper.
2518 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2519 from Martin Jerabek that uses macro names for all relevant character and
2520 string constants.
2522 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2523 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2524 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2525 these, but not everybody uses configure.
2527 19. A conditional group that had only one branch was not being correctly
2528 recognized as an item that could match an empty string. This meant that an
2529 enclosing group might also not be so recognized, causing infinite looping
2530 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2531 with the subject "ab", where knowledge that the repeated group can match
2532 nothing is needed in order to break the loop.
2534 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2535 exec(), but without supplying a callout function, matching went wrong.
2537 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2538 leak if the size of the offset vector was greater than 30. When the vector
2539 is smaller, the saved offsets during recursion go onto a local stack
2540 vector, but for larger vectors malloc() is used. It was failing to free
2541 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2542 error, in fact).
2544 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2545 heapframe that is used only when UTF-8 support is enabled. This caused no
2546 problem, but was untidy.
2548 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2549 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2550 included within another project.
2552 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2553 slightly modified by me:
2555 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2556 not building pcregrep.
2558 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2559 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2561 25. Forward references, both numeric and by name, in patterns that made use of
2562 duplicate group numbers, could behave incorrectly or give incorrect errors,
2563 because when scanning forward to find the reference group, PCRE was not
2564 taking into account the duplicate group numbers. A pattern such as
2565 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2567 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2568 the feature of a custom pointer more persuasive (as requested by a user).
2570 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2571 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2572 pre-defined.
2574 28. Added support for (*UTF8) at the start of a pattern.
2576 29. Arrange for flags added by the "release type" setting in CMake to be shown
2577 in the configuration summary.
2580 Version 7.8 05-Sep-08
2581 ---------------------
2583 1. Replaced UCP searching code with optimized version as implemented for Ad
2584 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2585 stage table and inline lookup instead of a function, giving speed ups of 2
2586 to 5 times on some simple patterns that I tested. Permission was given to
2587 distribute the MultiStage2.py script that generates the tables (it's not in
2588 the tarball, but is in the Subversion repository).
2590 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2591 scripts.
2593 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2594 a group with a zero qualifier. The result of the study could be incorrect,
2595 or the function might crash, depending on the pattern.
2597 4. Caseless matching was not working for non-ASCII characters in back
2598 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2599 It now works when Unicode Property Support is available.
2601 5. In pcretest, an escape such as \x{de} in the data was always generating
2602 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2603 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2604 truncation.
2606 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2608 7. Added two (int) casts to pcregrep when printing the difference of two
2609 pointers, in case they are 64-bit values.
2611 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2612 test 2 if it fails.
2614 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2615 and a #define of that name to empty if it is not externally set. This is to
2616 allow users of MSVC to set it if necessary.
2618 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2619 the convenience functions in the pcre_get.c source file.
2621 11. An option change at the start of a pattern that had top-level alternatives
2622 could cause overwriting and/or a crash. This command provoked a crash in
2623 some environments:
2625 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2627 This potential security problem was recorded as CVE-2008-2371.
2629 12. For a pattern where the match had to start at the beginning or immediately
2630 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2631 pcre_dfa_exec() could read past the end of the passed subject if there was
2632 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2633 pcretest so that it places the subject at the end of its malloc-ed buffer.
2635 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2636 exec() might read past the end of the data buffer in UTF-8 mode.
2638 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2639 the data contained the byte 0x85 as part of a UTF-8 character within its
2640 first line. This applied both to normal and DFA matching.
2642 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2643 /^[^d]*?$/8 failed to match "abc".
2645 16. Added a missing copyright notice to pcrecpp_internal.h.
2647 17. Make it more clear in the documentation that values returned from
2648 pcre_exec() in ovector are byte offsets, not character counts.
2650 18. Tidied a few places to stop certain compilers from issuing warnings.
2652 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2653 supplied by Stefan Weber. I made a further small update for 7.8 because
2654 there is a change of source arrangements: the pcre_searchfuncs.c module is
2655 replaced by pcre_ucd.c.
2658 Version 7.7 07-May-08
2659 ---------------------
2661 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2662 a string to a long long, pretend we don't even have a long long." This is
2663 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2665 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2666 pre-7.6 versions, which defined a global no_arg variable instead of putting
2667 it in the RE class. (See also #8 below.)
2669 3. Remove a line of dead code, identified by coverity and reported by Nuno
2670 Lopes.
2672 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2674 (1) The include/exclude patterns were being applied to the whole pathnames
2675 of files, instead of just to the final components.
2677 (2) If there was more than one level of directory, the subdirectories were
2678 skipped unless they satisfied the include/exclude conditions. This is
2679 inconsistent with GNU grep (and could even be seen as contrary to the
2680 pcregrep specification - which I improved to make it absolutely clear).
2681 The action now is always to scan all levels of directory, and just
2682 apply the include/exclude patterns to regular files.
2684 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2685 --exclude_dir in the tests to avoid scanning .svn directories.
2687 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2688 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2689 doesn't support NULs in patterns.
2691 7. Added some missing "const"s to declarations of static tables in
2692 pcre_compile.c and pcre_dfa_exec.c.
2694 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2695 caused by fix #2 above. (Subsequently also a second patch to fix the
2696 first patch. And a third patch - this was a messy problem.)
2698 9. Applied Craig's patch to remove the use of push_back().
2700 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2701 matching function regexec().
2703 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2704 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2705 references. PCRE supports relative numbers with this syntax (I don't think
2706 Oniguruma does).
2708 12. Previously, a group with a zero repeat such as (...){0} was completely
2709 omitted from the compiled regex. However, this means that if the group
2710 was called as a subroutine from elsewhere in the pattern, things went wrong
2711 (an internal error was given). Such groups are now left in the compiled
2712 pattern, with a new opcode that causes them to be skipped at execution
2713 time.
2715 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2716 to the way PCRE behaves:
2718 (a) A lone ] character is dis-allowed (Perl treats it as data).
2720 (b) A back reference to an unmatched subpattern matches an empty string
2721 (Perl fails the current match path).
2723 (c) A data ] in a character class must be notated as \] because if the
2724 first data character in a class is ], it defines an empty class. (In
2725 Perl it is not possible to have an empty class.) The empty class []
2726 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2727 The negative empty class [^] matches any one character, independently
2728 of the DOTALL setting.
2730 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2731 non-existent subpattern following a character class starting with ']' and
2732 containing () gave an internal compiling error instead of "reference to
2733 non-existent subpattern". Fortunately, when the pattern did exist, the
2734 compiled code was correct. (When scanning forwards to check for the
2735 existence of the subpattern, it was treating the data ']' as terminating
2736 the class, so got the count wrong. When actually compiling, the reference
2737 was subsequently set up correctly.)
2739 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2740 it was being rejected as not supported by pcre_dfa_exec(), even though
2741 other assertions are supported. I have made pcre_dfa_exec() support
2742 (*FAIL).
2744 16. The implementation of 13c above involved the invention of a new opcode,
2745 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2746 cannot be changed at match time, I realized I could make a small
2747 improvement to matching performance by compiling OP_ALLANY instead of
2748 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2749 on the OP_ANY path.
2751 17. Compiling pcretest on Windows with readline support failed without the
2752 following two fixes: (1) Make the unistd.h include conditional on
2753 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2755 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2756 ncurses library to be included for pcretest when ReadLine support is
2757 requested, but also to allow for it to be overridden. This patch came from
2758 Daniel Bergström.
2760 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2761 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2762 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2763 spotting this.
2766 Version 7.6 28-Jan-08
2767 ---------------------
2769 1. A character class containing a very large number of characters with
2770 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2771 overflow.
2773 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2774 HAVE_LONG_LONG is not defined.
2776 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2777 bring them up to date and include new features. This patch includes:
2779 - Fixed PH's badly added libz and libbz2 support.
2780 - Fixed a problem with static linking.
2781 - Added pcredemo. [But later removed - see 7 below.]
2782 - Fixed dftables problem and added an option.
2783 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2785 - Added readline support for pcretest.
2786 - Added an listing of the option settings after cmake has run.
2788 4. A user submitted a patch to Makefile that makes it easy to create
2789 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2790 Makefile.am that cause it to include this special target, without
2791 affecting anything else. Note that the same mingw target plus all
2792 the other distribution libraries and programs are now supported
2793 when configuring with CMake (see 6 below) instead of with
2794 Configure/Make.
2796 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2797 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2798 exported in the Windows port". It has not yet been confirmed that the patch
2799 solves the problem, but it does no harm.
2801 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2802 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2803 with CMake, and also correct the comment about stack recursion.
2805 7. Remove the automatic building of pcredemo from the ./configure system and
2806 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2807 of a program that users should build themselves after PCRE is installed, so
2808 building it automatically is not really right. What is more, it gave
2809 trouble in some build environments.
2811 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2814 Version 7.5 10-Jan-08
2815 ---------------------
2817 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2818 values in parens when parsing an RE using the C++ wrapper."
2820 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2821 Characters greater than 255 were excluded from the class instead of being
2822 included.
2824 3. The same bug as (2) above applied to negated POSIX classes such as
2825 [:^space:].
2827 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2828 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2829 I have changed it.
2831 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2832 first named subpattern) and a construct such as (?&a) would reference the
2833 first named subpattern whose name started with "a" (in other words, the
2834 length check was missing). Both these problems are fixed. "Subpattern name
2835 expected" is now given for (?&) (a zero-length name), and this patch also
2836 makes it give the same error for \k'' (previously it complained that that
2837 was a reference to a non-existent subpattern).
2839 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2840 this is right because (?- can be followed by option settings as well as by
2841 digits. I have, however, made the messages clearer.
2843 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2844 than the number used in the conditional) now cause a compile-time error.
2845 This is actually not compatible with Perl, which accepts such patterns, but
2846 treats the conditional as always being FALSE (as PCRE used to), but it
2847 seems to me that giving a diagnostic is better.
2849 8. Change "alphameric" to the more common word "alphanumeric" in comments
2850 and messages.
2852 9. Fix two occurrences of "backslash" in comments that should have been
2853 "backspace".
2855 10. Remove two redundant lines of code that can never be obeyed (their function
2856 was moved elsewhere).
2858 11. The program that makes PCRE's Unicode character property table had a bug
2859 which caused it to generate incorrect table entries for sequences of
2860 characters that have the same character type, but are in different scripts.
2861 It amalgamated them into a single range, with the script of the first of
2862 them. In other words, some characters were in the wrong script. There were
2863 thirteen such cases, affecting characters in the following ranges:
2865 U+002b0 - U+002c1
2866 U+0060c - U+0060d
2867 U+0061e - U+00612
2868 U+0064b - U+0065e
2869 U+0074d - U+0076d
2870 U+01800 - U+01805
2871 U+01d00 - U+01d77
2872 U+01d9b - U+01dbf
2873 U+0200b - U+0200f
2874 U+030fc - U+030fe
2875 U+03260 - U+0327f
2876 U+0fb46 - U+0fbb1
2877 U+10450 - U+1049d
2879 12. The -o option (show only the matching part of a line) for pcregrep was not
2880 compatible with GNU grep in that, if there was more than one match in a
2881 line, it showed only the first of them. It now behaves in the same way as
2882 GNU grep.
2884 13. If the -o and -v options were combined for pcregrep, it printed a blank
2885 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2886 does the same. The return code can be used to tell if there were any
2887 non-matching lines.
2889 14. Added --file-offsets and --line-offsets to pcregrep.
2891 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2892 infinitely looping recursion. The bug was that positive lookaheads were not
2893 being skipped when checking for a possible empty match (negative lookaheads
2894 and both kinds of lookbehind were skipped).
2896 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2897 inclusion of <windows.h> to before rather than after the definition of
2898 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2900 17. Specifying a possessive quantifier with a specific limit for a Unicode
2901 character property caused pcre_compile() to compile bad code, which led at
2902 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2903 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2904 caused the error; without that there was no problem.
2906 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2908 19. Added --enable-pcretest-libreadline.
2910 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2911 RE::GlobalReplace(). As a result, the number of replacements returned was
2912 double what it should be. I removed one of the increments, but Craig sent a
2913 later patch that removed the other one (the right fix) and added unit tests
2914 that check the return values (which was not done before).
2916 21. Several CMake things:
2918 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2919 the names libpcre and libpcreposix, not just pcre and pcreposix.
2921 (2) The above change means that pcretest and pcregrep are now correctly
2922 linked with the newly-built libraries, not previously installed ones.
2926 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2927 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2928 UTF-8 newline character). The key issue is that the pattern starts .*;
2929 this means that the match must be either at the beginning, or after a
2930 newline. The bug was in the code for advancing after a failed match and
2931 checking that the new position followed a newline. It was not taking
2932 account of UTF-8 characters correctly.
2934 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2935 character classes. PCRE was not treating the sequence [:...:] as a
2936 character class unless the ... were all letters. Perl, however, seems to
2937 allow any characters between [: and :], though of course it rejects as
2938 unknown any "names" that contain non-letters, because all the known class
2939 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2940 for example, whereas PCRE did not - it did not recognize a POSIX character
2941 class. This seemed a bit dangerous, so the code has been changed to be
2942 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2943 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2944 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2945 Perl does, and where it didn't before.
2947 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2948 Windows environments %n is disabled by default.
2951 Version 7.4 21-Sep-07
2952 ---------------------
2954 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2955 means that a class such as [\s] counted as "explicit reference to CR or
2956 LF". That isn't really right - the whole point of the change was to try to
2957 help when there was an actual mention of one of the two characters. So now
2958 the change happens only if \r or \n (or a literal CR or LF) character is
2959 encountered.
2961 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2962 of both had grown to the point where there were only 3 bits left.
2963 Fortunately, there was spare space in the data structure, and so I have
2964 moved the internal flags into a new 16-bit field to free up more option
2965 bits.
2967 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2968 but did not set the internal JCHANGED flag - either of these is enough to
2969 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2970 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2971 start sets both bits.
2973 4. Added options (at build time, compile time, exec time) to change \R from
2974 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2976 5. doc/pcresyntax.html was missing from the distribution.
2978 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2979 compatibility, even though it is no longer used.
2981 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2982 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2983 windows.h file is present (where different names are used). [This was
2984 reversed later after testing - see 16 below.]
2986 8. Changed all #include <config.h> to #include "config.h". There were also
2987 some further <pcre.h> cases that I changed to "pcre.h".
2989 9. When pcregrep was used with the --colour option, it missed the line ending
2990 sequence off the lines that it output.
2992 10. It was pointed out to me that arrays of string pointers cause lots of
2993 relocations when a shared library is dynamically loaded. A technique of
2994 using a single long string with a table of offsets can drastically reduce
2995 these. I have refactored PCRE in four places to do this. The result is
2996 dramatic:
2998 Originally: 290
2999 After changing UCP table: 187
3000 After changing error message table: 43
3001 After changing table of "verbs" 36
3002 After changing table of Posix names 22
3004 Thanks to the folks working on Gregex for glib for this insight.
3006 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3007 unicode-properties was also set.
3009 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3011 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3012 checked only for CRLF.
3014 14. Added casts to pcretest.c to avoid compiler warnings.
3016 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3018 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3019 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3020 entirely. This removes changes made in 7 above.
3022 17. The CMake files have been updated, and there is now more information about
3023 building with CMake in the NON-UNIX-USE document.
3026 Version 7.3 28-Aug-07
3027 ---------------------
3029 1. In the rejigging of the build system that eventually resulted in 7.1, the
3030 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3031 brackets there is not right, since it causes compilers to look for an
3032 installed pcre.h, not the version that is in the source that is being
3033 compiled (which of course may be different). I have changed it back to:
3035 #include "pcre.h"
3037 I have a vague recollection that the change was concerned with compiling in
3038 different directories, but in the new build system, that is taken care of
3039 by the VPATH setting the Makefile.
3041 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3042 when the subject happened to end in the byte 0x85 (e.g. if the last
3043 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3044 characters but of course it shouldn't be taken as a newline when it is part
3045 of another character. The bug was that, for an unlimited repeat of . in
3046 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3047 characters when looking for a newline.
3049 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3051 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3052 in debug output.
3054 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3055 long printing in the pcrecpp unittest when running under MinGW.
3057 6. ESC_K was left out of the EBCDIC table.
3059 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3060 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3061 limit also applies to "virtual nesting" when a pattern is recursive, and in
3062 this case 1000 isn't so big. I have been able to remove this limit at the
3063 expense of backing off one optimization in certain circumstances. Normally,
3064 when pcre_exec() would call its internal match() function recursively and
3065 immediately return the result unconditionally, it uses a "tail recursion"
3066 feature to save stack. However, when a subpattern that can match an empty
3067 string has an unlimited repetition quantifier, it no longer makes this
3068 optimization. That gives it a stack frame in which to save the data for
3069 checking that an empty string has been matched. Previously this was taken
3070 from the 1000-entry workspace that had been reserved. So now there is no
3071 explicit limit, but more stack is used.
3073 8. Applied Daniel's patches to solve problems with the import/export magic
3074 syntax that is required for Windows, and which was going wrong for the
3075 pcreposix and pcrecpp parts of the library. These were overlooked when this
3076 problem was solved for the main library.
3078 9. There were some crude static tests to avoid integer overflow when computing
3079 the size of patterns that contain repeated groups with explicit upper
3080 limits. As the maximum quantifier is 65535, the maximum group length was
3081 set at 30,000 so that the product of these two numbers did not overflow a
3082 32-bit integer. However, it turns out that people want to use groups that
3083 are longer than 30,000 bytes (though not repeat them that many times).
3084 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3085 made it possible to implement the integer overflow checks in a much more
3086 dynamic way, which I have now done. The artificial limitation on group
3087 length has been removed - we now have only the limit on the total length of
3088 the compiled pattern, which depends on the LINK_SIZE setting.
3090 10. Fixed a bug in the documentation for get/copy named substring when
3091 duplicate names are permitted. If none of the named substrings are set, the
3092 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3093 empty string.
3095 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3096 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3097 because the ] is interpreted as the first data character and the
3098 terminating ] is not found. PCRE has been made compatible with Perl in this
3099 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3100 cause memory overwriting.
3102 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3103 string has been matched (to stop an infinite loop). It was not recognizing
3104 a conditional subpattern that could match an empty string if that
3105 subpattern was within another subpattern. For example, it looped when
3106 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3107 condition was not nested. This bug has been fixed.
3109 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3110 past the start of the subject in the presence of bytes with the top bit
3111 set, for example "\x8aBCD".
3113 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3114 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3116 14. Optimized (?!) to (*FAIL).
3118 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3119 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3120 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3121 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3122 does: it's just the validity check that is more restrictive.
3124 16. Inserted checks for integer overflows during escape sequence (backslash)
3125 processing, and also fixed erroneous offset values for syntax errors during
3126 backslash processing.
3128 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3129 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3131 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3132 caused an overrun.
3134 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3135 something other than just ASCII characters) inside a group that had an
3136 unlimited repeat caused a loop at compile time (while checking to see
3137 whether the group could match an empty string).
3139 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3140 [\P{Any}] did so. (Error in the code for printing property names.)
3142 21. An orphan \E inside a character class could cause a crash.
3144 22. A repeated capturing bracket such as (A)? could cause a wild memory
3145 reference during compilation.
3147 23. There are several functions in pcre_compile() that scan along a compiled
3148 expression for various reasons (e.g. to see if it's fixed length for look
3149 behind). There were bugs in these functions when a repeated \p or \P was
3150 present in the pattern. These operators have additional parameters compared
3151 with \d, etc, and these were not being taken into account when moving along
3152 the compiled data. Specifically:
3154 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3155 length.
3157 (b) An item such as \pL+ within a repeated group could cause crashes or
3158 loops.
3160 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3161 "reference to non-existent subpattern" error.
3163 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3165 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3166 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3168 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3170 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3171 character were causing crashes (broken optimization).
3173 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3174 \p or \P) caused a compile-time loop.
3176 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3177 break. For example, the unstudied pattern [\r\n]A does not match the string
3178 "\r\nA" because change 7.0/46 below moves the current point on by two
3179 characters after failing to match at the start. However, the pattern \nA
3180 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3181 the same is true. There doesn't seem any very clean way out of this, but
3182 what I have chosen to do makes the common cases work: PCRE now takes note
3183 of whether there can be an explicit match for \r or \n anywhere in the
3184 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3185 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3186 pattern has explicit CR or LF references.
3188 29. Added (*CR) etc for changing newline setting at start of pattern.
3191 Version 7.2 19-Jun-07
3192 ---------------------
3194 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3195 which is apparently normally available under Windows.
3197 2. Re-jig the pcregrep tests with different newline settings in an attempt
3198 to make them independent of the local environment's newline setting.
3200 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3202 4. Some of the "internals" tests were previously cut out when the link size
3203 was not 2, because the output contained actual offsets. The recent new
3204 "Z" feature of pcretest means that these can be cut out, making the tests
3205 usable with all link sizes.
3207 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3208 stack recursion. This gives a massive performance boost under BSD, but just
3209 a small improvement under Linux. However, it saves one field in the frame
3210 in all cases.
3212 6. Added more features from the forthcoming Perl 5.10:
3214 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3215 recursion call. It refers to the nth most recently opened parentheses.
3217 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3218 to be opened parentheses.
3220 (c) Conditions that refer to capturing parentheses can be specified
3221 relatively, for example, (?(-2)... or (?(+3)...
3223 (d) \K resets the start of the current match so that everything before
3224 is not part of it.
3226 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3228 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3229 reference syntax.
3231 (g) (?| introduces a group in which the numbering of parentheses in each
3232 alternative starts with the same number.
3234 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3236 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3239 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3240 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3241 for detecting groups that can match an empty string.
3243 9. A pattern with a very large number of alternatives (more than several
3244 hundred) was running out of internal workspace during the pre-compile
3245 phase, where pcre_compile() figures out how much memory will be needed. A
3246 bit of new cunning has reduced the workspace needed for groups with
3247 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3248 workspace instead of running out of the 4096 that are available.
3250 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3252 11. Applied patch from Google to remove an optimization that didn't quite work.
3253 The report of the bug said:
3255 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3256 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3257 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3259 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3260 it matched the wrong number of bytes.
3263 Version 7.1 24-Apr-07
3264 ---------------------
3266 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3267 that is more "standard", making use of automake and other Autotools. There
3268 is some re-arrangement of the files and adjustment of comments consequent
3269 on this.
3271 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3272 for recursive directory scanning broke on some systems because the files
3273 are not scanned in any specific order and on different systems the order
3274 was different. A call to "sort" has been inserted into RunGrepTest for the
3275 approprate test as a short-term fix. In the longer term there may be an
3276 alternative.
3278 3. I had an email from Eric Raymond about problems translating some of PCRE's
3279 man pages to HTML (despite the fact that I distribute HTML pages, some
3280 people do their own conversions for various reasons). The problems
3281 concerned the use of low-level troff macros .br and .in. I have therefore
3282 removed all such uses from the man pages (some were redundant, some could
3283 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3284 HTML has been updated to handle .nf/.fi and to complain if it encounters
3285 .br or .in.
3287 4. Updated comments in configure.ac that get placed in config.h.in and also
3288 arranged for config.h to be included in the distribution, with the name
3289 config.h.generic, for the benefit of those who have to compile without
3290 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3292 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3293 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3294 makevp.bat for the current PCRE, using the additional files
3295 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3297 6. A Windows user reported a minor discrepancy with test 2, which turned out
3298 to be caused by a trailing space on an input line that had got lost in his
3299 copy. The trailing space was an accident, so I've just removed it.
3301 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3302 that is needed.
3304 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3305 as "const" (a) because they are and (b) because it helps the PHP
3306 maintainers who have recently made a script to detect big data structures
3307 in the php code that should be moved to the .rodata section. I remembered
3308 to update Builducptable as well, so it won't revert if ucptable.h is ever
3309 re-created.
3311 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3312 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3313 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3314 support is not required. This saves 1.5-2K of code, which is important in
3315 some applications.
3317 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3318 so as not to refer to the tables, even though these functions will never be
3319 called when UTF-8 support is disabled. Otherwise there are problems with a
3320 shared library.
3322 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3324 (a) It was defining its arguments as char * instead of void *.
3326 (b) It was assuming that all moves were upwards in memory; this was true
3327 a long time ago when I wrote it, but is no longer the case.
3329 The emulated memove() is provided for those environments that have neither
3330 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3331 is clearly not the case, as these two bugs were recently reported.
3333 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3334 and Detrail to create the HTML documentation, the .txt form of the man
3335 pages, and it removes trailing spaces from listed files. It also creates
3336 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3337 case, it wraps all the #defines with #ifndefs. This script should be run
3338 before "make dist".
3340 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3341 with Unicode property support.
3343 (a) For a maximizing quantifier, if the two different cases of the
3344 character were of different lengths in their UTF-8 codings (there are
3345 some cases like this - I found 11), and the matching function had to
3346 back up over a mixture of the two cases, it incorrectly assumed they
3347 were both the same length.
3349 (b) When PCRE was configured to use the heap rather than the stack for
3350 recursion during matching, it was not correctly preserving the data for
3351 the other case of a UTF-8 character when checking ahead for a match
3352 while processing a minimizing repeat. If the check also involved
3353 matching a wide character, but failed, corruption could cause an
3354 erroneous result when trying to check for a repeat of the original
3355 character.
3357 13. Some tidying changes to the testing mechanism:
3359 (a) The RunTest script now detects the internal link size and whether there
3360 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3361 values substituted by "configure". (The RunGrepTest script already did
3362 this for UTF-8.) The configure.ac script no longer substitutes the
3363 relevant variables.
3365 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3366 with length and offset values. This means that the output is different
3367 for different internal link sizes. Test 2 is skipped for link sizes
3368 other than 2 because of this, bypassing the problem. Unfortunately,
3369 there was also a test in test 3 (the locale tests) that used /B and
3370 failed for link sizes other than 2. Rather than cut the whole test out,
3371 I have added a new /Z option to pcretest that replaces the length and
3372 offset values with spaces. This is now used to make test 3 independent
3373 of link size. (Test 2 will be tidied up later.)
3375 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3376 segmentation fault instead of returning the appropriate error message.
3378 15. In multiline mode when the newline sequence was set to "any", the pattern
3379 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3380 This doesn't seem right; it now treats the CRLF combination as the line
3381 ending, and so does not match in that case. It's only a pattern such as ^$
3382 that would hit this one: something like ^ABC$ would have failed after \r
3383 and then tried again after \r\n.
3385 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3386 in an attempt to make files that differ only in their line terminators
3387 compare equal. This works on Linux.
3389 17. Under certain error circumstances pcregrep might try to free random memory
3390 as it exited. This is now fixed, thanks to valgrind.
3392 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3393 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3394 was because its rules for how to advance for /g after matching an empty
3395 string at the end of a line did not allow for this case. They now check for
3396 it specially.
3398 20. pcretest is supposed to handle patterns and data of any length, by
3399 extending its buffers when necessary. It was getting this wrong when the
3400 buffer for a data line had to be extended.
3402 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3403 CRLF as a newline sequence.
3405 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3406 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3407 I have nevertheless tidied it up.
3409 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3411 24. Added a man page for pcre-config.
3414 Version 7.0 19-Dec-06
3415 ---------------------
3417 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3418 moving to gcc 4.1.1.
3420 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3421 sys/time.h, which is documented as needed for this function. It doesn't
3422 seem to matter on Linux, but it showed up on some releases of OS X.
3424 3. It seems that there are systems where bytes whose values are greater than
3425 127 match isprint() in the "C" locale. The "C" locale should be the
3426 default when a C program starts up. In most systems, only ASCII printing
3427 characters match isprint(). This difference caused the output from pcretest
3428 to vary, making some of the tests fail. I have changed pcretest so that:
3430 (a) When it is outputting text in the compiled version of a pattern, bytes
3431 other than 32-126 are always shown as hex escapes.
3433 (b) When it is outputting text that is a matched part of a subject string,
3434 it does the same, unless a different locale has been set for the match
3435 (using the /L modifier). In this case, it uses isprint() to decide.
3437 4. Fixed a major bug that caused incorrect computation of the amount of memory
3438 required for a compiled pattern when options that changed within the
3439 pattern affected the logic of the preliminary scan that determines the
3440 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3441 that the computed length was too small. The symptoms of this bug were
3442 either the PCRE error "internal error: code overflow" from pcre_compile(),
3443 or a glibc crash with a message such as "pcretest: free(): invalid next
3444 size (fast)". Examples of patterns that provoked this bug (shown in
3445 pcretest format) are:
3447 /(?-x: )/x
3448 /(?x)(?-x: \s*#\s*)/
3449 /((?i)[\x{c0}])/8
3450 /(?i:[\x{c0}])/8
3452 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3453 is now done differently.
3455 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3456 wrapper classes; (b) implement a new function in the C++ scanner that is
3457 more efficient than the old way of doing things because it avoids levels of
3458 recursion in the regex matching; (c) add a paragraph to the documentation
3459 for the FullMatch() function.
3461 6. The escape sequence \n was being treated as whatever was defined as
3462 "newline". Not only was this contrary to the documentation, which states
3463 that \n is character 10 (hex 0A), but it also went horribly wrong when
3464 "newline" was defined as CRLF. This has been fixed.
3466 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3467 was being set to -1 for the "end of line" case (supposedly a value that no
3468 character can have). Though this value is never used (the check for end of
3469 line is "zero bytes in current character"), it caused compiler complaints.
3470 I've changed it to 0xffffffff.
3472 8. In pcre_version.c, the version string was being built by a sequence of
3473 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3474 string (as it is for production releases) called a macro with an empty
3475 argument. The C standard says the result of this is undefined. The gcc
3476 compiler treats it as an empty string (which was what was wanted) but it is
3477 reported that Visual C gives an error. The source has been hacked around to
3478 avoid this problem.
3480 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3481 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3482 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3483 of them did).
3485 10. Originally, pcretest opened its input and output without "b"; then I was
3486 told that "b" was needed in some environments, so it was added for release
3487 5.0 to both the input and output. (It makes no difference on Unix-like
3488 systems.) Later I was told that it is wrong for the input on Windows. I've
3489 now abstracted the modes into two macros, to make it easier to fiddle with
3490 them, and removed "b" from the input mode under Windows.
3492 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3494 12. Added -help and --help to pcretest as an official way of being reminded
3495 of the options.
3497 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3498 and pcrecpp.cc because they annoy compilers at high warning levels.
3500 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3502 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3503 scripts are not C programs :-) and which was not noticed because it works
3504 on Linux.
3506 16. pcretest is supposed to handle any length of pattern and data line (as one
3507 line or as a continued sequence of lines) by extending its input buffer if
3508 necessary. This feature was broken for very long pattern lines, leading to
3509 a string of junk being passed to pcre_compile() if the pattern was longer
3510 than about 50K.
3512 17. I have done a major re-factoring of the way pcre_compile() computes the
3513 amount of memory needed for a compiled pattern. Previously, there was code
3514 that made a preliminary scan of the pattern in order to do this. That was
3515 OK when PCRE was new, but as the facilities have expanded, it has become
3516 harder and harder to keep it in step with the real compile phase, and there
3517 have been a number of bugs (see for example, 4 above). I have now found a
3518 cunning way of running the real compile function in a "fake" mode that
3519 enables it to compute how much memory it would need, while actually only
3520 ever using a few hundred bytes of working memory and without too many
3521 tests of the mode. This should make future maintenance and development
3522 easier. A side effect of this work is that the limit of 200 on the nesting
3523 depth of parentheses has been removed (though this was never a serious
3524 limitation, I suspect). However, there is a downside: pcre_compile() now
3525 runs more slowly than before (30% or more, depending on the pattern). I
3526 hope this isn't a big issue. There is no effect on runtime performance.
3528 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3529 newline (only possible for the last line of a file) and it was a
3530 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3532 19. Added additional timing features to pcretest. (1) The -tm option now times
3533 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3534 separate command line item, by a number that specifies the number of
3535 repeats to use when timing. The default is 50000; this gives better
3536 precision, but takes uncomfortably long for very large patterns.
3538 20. Extended pcre_study() to be more clever in cases where a branch of a
3539 subpattern has no definite first character. For example, (a*|b*)[cd] would
3540 previously give no result from pcre_study(). Now it recognizes that the
3541 first character must be a, b, c, or d.
3543 21. There was an incorrect error "recursive call could loop indefinitely" if
3544 a subpattern (or the entire pattern) that was being tested for matching an
3545 empty string contained only one non-empty item after a nested subpattern.
3546 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3547 incorrectly, because the \d was being skipped in the check.
3549 22. The pcretest program now has a new pattern option /B and a command line
3550 option -b, which is equivalent to adding /B to every pattern. This causes
3551 it to show the compiled bytecode, without the additional information that
3552 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3553 is the same as /B/I).
3555 23. A new optimization is now able automatically to treat some sequences such
3556 as a*b as a*+b. More specifically, if something simple (such as a character
3557 or a simple class like \d) has an unlimited quantifier, and is followed by
3558 something that cannot possibly match the quantified thing, the quantifier
3559 is automatically "possessified".
3561 24. A recursive reference to a subpattern whose number was greater than 39
3562 went wrong under certain circumstances in UTF-8 mode. This bug could also
3563 have affected the operation of pcre_study().
3565 25. Realized that a little bit of performance could be had by replacing
3566 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3568 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3570 27. Possessive quantifiers such as a++ were previously implemented by turning
3571 them into atomic groups such as ($>a+). Now they have their own opcodes,
3572 which improves performance. This includes the automatically created ones
3573 from 23 above.
3575 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3576 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3577 the first matched character to be a colon. This applied both to named and
3578 numbered groups.
3580 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3582 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3583 building PCRE on Windows easier, so I have included it in the distribution.
3585 31. There is now a check in pcretest against a ridiculously large number being
3586 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3587 loop, the loop is abandoned.
3589 32. Forward references to subpatterns in conditions such as (?(2)...) where
3590 subpattern 2 is defined later cause pcre_compile() to search forwards in
3591 the pattern for the relevant set of parentheses. This search went wrong
3592 when there were unescaped parentheses in a character class, parentheses
3593 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3595 33. "Subroutine" calls and backreferences were previously restricted to
3596 referencing subpatterns earlier in the regex. This restriction has now
3597 been removed.
3599 34. Added a number of extra features that are going to be in Perl 5.10. On the
3600 whole, these are just syntactic alternatives for features that PCRE had
3601 previously implemented using the Python syntax or my own invention. The
3602 other formats are all retained for compatibility.
3604 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3605 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3606 also .NET compatible.
3608 (b) A recursion or subroutine call to a named group can now be defined as
3609 (?&name) as well as (?P>name).
3611 (c) A backreference to a named group can now be defined as \k<name> or
3612 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3613 5.10, are also .NET compatible.
3615 (d) A conditional reference to a named group can now use the syntax
3616 (?(<name>) or (?('name') as well as (?(name).
3618 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3619 groups (named and numbered) that are never evaluated inline, but can be
3620 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3621 is always false. There may be only one alternative in such a group.
3623 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3624 as the simple (?(R). The condition is true only if the most recent
3625 recursion is that of the given number or name. It does not search out
3626 through the entire recursion stack.
3628 (g) The escape \gN or \g{N} has been added, where N is a positive or
3629 negative number, specifying an absolute or relative reference.
3631 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3632 some "unreachable code" warnings.
3634 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3635 things, this adds five new scripts.
3637 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3638 There were also incompatibilities regarding the handling of \Q..\E inside
3639 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3640 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3642 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3643 matches an empty string, and forcibly breaks the loop. There were bugs in
3644 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3645 against aaaa the result was just "a" rather than "aaaa", for example. Two
3646 separate and independent bugs (that affected different cases) have been
3647 fixed.
3649 39. Refactored the code to abolish the use of different opcodes for small
3650 capturing bracket numbers. This is a tidy that I avoided doing when I
3651 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3652 The new approach is not only tidier, it makes it possible to reduce the
3653 memory needed to fix the previous bug (38).
3655 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3656 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3657 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3658 mode.
3660 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3661 report.
3663 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3664 copy construction and assignment in the C++ wrapper.
3666 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3667 couple of bugs that could have given wrong results in the "--newline=crlf"
3668 case.
3670 44. Added a number of casts and did some reorganization of signed/unsigned int
3671 variables following suggestions from Dair Grant. Also renamed the variable
3672 "this" as "item" because it is a C++ keyword.
3674 45. Arranged for dftables to add
3676 #include "pcre_internal.h"
3678 to pcre_chartables.c because without it, gcc 4.x may remove the array
3679 definition from the final binary if PCRE is built into a static library and
3680 dead code stripping is activated.
3682 46. For an unanchored pattern, if a match attempt fails at the start of a
3683 newline sequence, and the newline setting is CRLF or ANY, and the next two
3684 characters are CRLF, advance by two characters instead of one.
3687 Version 6.7 04-Jul-06
3688 ---------------------
3690 1. In order to handle tests when input lines are enormously long, pcretest has
3691 been re-factored so that it automatically extends its buffers when
3692 necessary. The code is crude, but this _is_ just a test program. The
3693 default size has been increased from 32K to 50K.
3695 2. The code in pcre_study() was using the value of the re argument before
3696 testing it for NULL. (Of course, in any sensible call of the function, it
3697 won't be NULL.)
3699 3. The memmove() emulation function in pcre_internal.h, which is used on
3700 systems that lack both memmove() and bcopy() - that is, hardly ever -
3701 was missing a "static" storage class specifier.
3703 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3704 containing an extended class (one that cannot be represented by a bitmap
3705 because it contains high-valued characters or Unicode property items, e.g.
3706 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3707 pattern, but PCRE should not loop if you do not (it no longer does).
3708 [Detail: two cases were found: (a) a repeated subpattern containing an
3709 extended class; (b) a recursive reference to a subpattern that followed a
3710 previous extended class. It wasn't skipping over the extended class
3711 correctly when UTF-8 mode was not set.]
3713 5. A negated single-character class was not being recognized as fixed-length
3714 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3715 compile error "lookbehind assertion is not fixed length".
3717 6. The RunPerlTest auxiliary script was showing an unexpected difference
3718 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3719 write a Perl script that can interpret lines of an input file either as
3720 byte characters or as UTF-8, which is what "perltest" was being required to
3721 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3722 can't do is switch easily at run time between having the "use utf8;" pragma
3723 or not. In the end, I fudged it by using the RunPerlTest script to insert
3724 "use utf8;" explicitly for the UTF-8 tests.
3726 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3727 the end of the subject string, contrary to the documentation and to what
3728 Perl does. This was true of both matching functions. Now it matches only at
3729 the start of the subject and immediately after *internal* newlines.
3731 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3732 a pointer to an int instead of a pointer to an unsigned long int. This
3733 caused problems on 64-bit systems.
3735 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3736 instance of the 'standard' template library not being so standard".
3738 10. There was no check on the number of named subpatterns nor the maximum
3739 length of a subpattern name. The product of these values is used to compute
3740 the size of the memory block for a compiled pattern. By supplying a very
3741 long subpattern name and a large number of named subpatterns, the size
3742 computation could be caused to overflow. This is now prevented by limiting
3743 the length of names to 32 characters, and the number of named subpatterns
3744 to 10,000.
3746 11. Subpatterns that are repeated with specific counts have to be replicated in
3747 the compiled pattern. The size of memory for this was computed from the
3748 length of the subpattern and the repeat count. The latter is limited to
3749 65535, but there was no limit on the former, meaning that integer overflow
3750 could in principle occur. The compiled length of a repeated subpattern is
3751 now limited to 30,000 bytes in order to prevent this.
3753 12. Added the optional facility to have named substrings with the same name.
3755 13. Added the ability to use a named substring as a condition, using the
3756 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3757 are numbers (not recommended). Forward references are permitted.
3759 14. Added forward references in named backreferences (if you see what I mean).
3761 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3762 pattern could run off the end of the subject. For example, the pattern
3763 "(?s)(.{1,5})"8 did this with the subject "ab".
3765 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3766 PCRE_CASELESS was set when matching characters that were quantified with ?
3767 or *.
3769 17. A character class other than a single negated character that had a minimum
3770 but no maximum quantifier - for example [ab]{6,} - was not handled
3771 correctly by pce_dfa_exec(). It would match only one character.
3773 18. A valid (though odd) pattern that looked like a POSIX character
3774 class but used an invalid character after [ (for example [[,abc,]]) caused
3775 pcre_compile() to give the error "Failed: internal error: code overflow" or
3776 in some cases to crash with a glibc free() error. This could even happen if
3777 the pattern terminated after [[ but there just happened to be a sequence of
3778 letters, a binary zero, and a closing ] in the memory that followed.
3780 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3781 over the years. Originally (before any Unicode support), just the bottom 8
3782 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3783 output from "man perlunicode" includes this:
3785 The regular expression compiler produces polymorphic opcodes. That
3786 is, the pattern adapts to the data and automatically switches to
3787 the Unicode character scheme when presented with Unicode data--or
3788 instead uses a traditional byte scheme when presented with byte
3789 data.
3791 Sadly, a wide octal escape does not cause a switch, and in a string with
3792 no other multibyte characters, these octal escapes are treated as before.
3793 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3794 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3795 Unicode string.
3797 I have not perpetrated such confusion in PCRE. Up till now, it took just
3798 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3799 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3800 translate to the appropriate multibyte character.
3802 29. Applied some refactoring to reduce the number of warnings from Microsoft
3803 and Borland compilers. This has included removing the fudge introduced
3804 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3805 a warning about an unused variable.
3807 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3808 characters since release 4.0, because Perl (from release 5.004) does not.
3809 [Or at least, is documented not to: some releases seem to be in conflict
3810 with the documentation.] However, when a pattern was studied with
3811 pcre_study() and all its branches started with \s, PCRE still included VT
3812 as a possible starting character. Of course, this did no harm; it just
3813 caused an unnecessary match attempt.
3815 22. Removed a now-redundant internal flag bit that recorded the fact that case
3816 dependency changed within the pattern. This was once needed for "required
3817 byte" processing, but is no longer used. This recovers a now-scarce options
3818 bit. Also moved the least significant internal flag bit to the most-
3819 significant bit of the word, which was not previously used (hangover from
3820 the days when it was an int rather than a uint) to free up another bit for
3821 the future.
3823 23. Added support for CRLF line endings as well as CR and LF. As well as the
3824 default being selectable at build time, it can now be changed at runtime
3825 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3826 specify that it is scanning data with non-default line endings.
3828 24. Changed the definition of CXXLINK to make it agree with the definition of
3829 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3831 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3832 recursions. This makes a big different to stack usage for some patterns.
3834 26. If a subpattern containing a named recursion or subroutine reference such
3835 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3836 the space required for the compiled pattern went wrong and gave too small a
3837 value. Depending on the environment, this could lead to "Failed: internal
3838 error: code overflow at offset 49" or "glibc detected double free or
3839 corruption" errors.
3841 27. Applied patches from Google (a) to support the new newline modes and (b) to
3842 advance over multibyte UTF-8 characters in GlobalReplace.
3844 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3845 difference for some implementation of PCRE in some Windows version.
3847 29. Added some extra testing facilities to pcretest:
3849 \q<number> in a data line sets the "match limit" value
3850 \Q<number> in a data line sets the "match recursion limt" value
3851 -S <number> sets the stack size, where <number> is in megabytes
3853 The -S option isn't available for Windows.
3856 Version 6.6 06-Feb-06
3857 ---------------------
3859 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3860 in pcreposix.h. I have copied the definition from pcre.h.
3862 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3863 because pcre.h is no longer a built file.
3865 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3866 not normally included in the compiled code.
3869 Version 6.5 01-Feb-06
3870 ---------------------
3872 1. When using the partial match feature with pcre_dfa_exec(), it was not
3873 anchoring the second and subsequent partial matches at the new starting
3874 point. This could lead to incorrect results. For example, with the pattern
3875 /1234/, partially matching against "123" and then "a4" gave a match.
3877 2. Changes to pcregrep:
3879 (a) All non-match returns from pcre_exec() were being treated as failures
3880 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3881 error message is output. Some extra information is given for the
3883 probably the only errors that are likely to be caused by users (by
3884 specifying a regex that has nested indefinite repeats, for instance).
3885 If there are more than 20 of these errors, pcregrep is abandoned.
3887 (b) A binary zero was treated as data while matching, but terminated the
3888 output line if it was written out. This has been fixed: binary zeroes
3889 are now no different to any other data bytes.
3891 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3892 used to set a locale for matching. The --locale=xxxx long option has
3893 been added (no short equivalent) to specify a locale explicitly on the
3894 pcregrep command, overriding the environment variables.
3896 (d) When -B was used with -n, some line numbers in the output were one less
3897 than they should have been.
3899 (e) Added the -o (--only-matching) option.
3901 (f) If -A or -C was used with -c (count only), some lines of context were
3902 accidentally printed for the final match.
3904 (g) Added the -H (--with-filename) option.
3906 (h) The combination of options -rh failed to suppress file names for files
3907 that were found from directory arguments.
3909 (i) Added the -D (--devices) and -d (--directories) options.
3911 (j) Added the -F (--fixed-strings) option.
3913 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3915 (l) Added the --colo(u)r option.
3917 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3918 is not present by default.
3920 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3921 items such as (?R) or (?1), when the recursion could match a number of
3922 alternatives. If it matched one of the alternatives, but subsequently,
3923 outside the recursion, there was a failure, the code tried to back up into
3924 the recursion. However, because of the way PCRE is implemented, this is not
3925 possible, and the result was an incorrect result from the match.
3927 In order to prevent this happening, the specification of recursion has
3928 been changed so that all such subpatterns are automatically treated as
3929 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3931 4. I had overlooked the fact that, in some locales, there are characters for
3932 which isalpha() is true but neither isupper() nor islower() are true. In
3933 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3934 and ordfeminine) are like this. This affected the treatment of \w and \W
3935 when they appeared in character classes, but not when they appeared outside
3936 a character class. The bit map for "word" characters is now created
3937 separately from the results of isalnum() instead of just taking it from the
3938 upper, lower, and digit maps. (Plus the underscore character, of course.)
3940 5. The above bug also affected the handling of POSIX character classes such as
3941 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3942 permanent tables. Instead, the bit maps for such a class were previously
3943 created as the appropriate unions of the upper, lower, and digit bitmaps.
3944 Now they are created by subtraction from the [[:word:]] class, which has
3945 its own bitmap.
3947 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3948 It is created by subtracting the vertical space characters (\x09, \x0a,
3949 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3950 subtraction was done in the overall bitmap for a character class, meaning
3951 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3952 be recognized. This bug has been fixed.
3954 7. Patches from the folks at Google:
3956 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3957 real life, but is still worth protecting against".
3959 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3960 regular expressions".
3962 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3963 have it.
3965 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3966 "configure" and the latter not, in order to fix a problem somebody had
3967 with compiling the Arg class on HP-UX.
3969 (e) Improve the error-handling of the C++ wrapper a little bit.
3971 (f) New tests for checking recursion limiting.
3973 8. The pcre_memmove() function, which is used only if the environment does not
3974 have a standard memmove() function (and is therefore rarely compiled),
3975 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3976 returning a result (though PCRE never actually uses the result).
3978 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3979 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3980 returned instead of calling malloc() with an overflowing number that would
3981 most likely cause subsequent chaos.
3983 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3985 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3986 with this option is matched, the nmatch and pmatch options of regexec() are
3987 ignored.
3989 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3990 provided in case anyone wants to the the POSIX interface with UTF-8
3991 strings.
3993 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3994 C++ linking (needed for some HP-UX environments).
3996 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3997 (unused parameter) and in the pcre_printint() function (omitted "default"
3998 switch label when the default is to do nothing).
4000 15. Added some code to make it possible, when PCRE is compiled as a C++
4001 library, to replace subject pointers for pcre_exec() with a smart pointer
4002 class, thus making it possible to process discontinuous strings.
4004 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
4005 much the same function. They were added by different people who were trying
4006 to make PCRE easy to compile on non-Unix systems. It has been suggested
4007 that PCRE_EXPORT be abolished now that there is more automatic apparatus
4008 for compiling on Windows systems. I have therefore replaced it with
4009 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
4010 defaults to "extern" for C or "extern C" for C++, which works fine on
4011 Unix-like systems. It is now possible to override the value of PCRE_DATA_
4012 SCOPE with something explicit in config.h. In addition:
4014 (a) pcreposix.h still had just "extern" instead of either of these macros;
4015 I have replaced it with PCRE_DATA_SCOPE.
4017 (b) Functions such as _pcre_xclass(), which are internal to the library,
4018 but external in the C sense, all had PCRE_EXPORT in their definitions.
4019 This is apparently wrong for the Windows case, so I have removed it.
4020 (It makes no difference on Unix-like systems.)
4022 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
4023 of recursive calls to match(). This is different to MATCH_LIMIT because
4024 that limits the total number of calls to match(), not all of which increase
4025 the depth of recursion. Limiting the recursion depth limits the amount of
4026 stack (or heap if NO_RECURSE is set) that is used. The default can be set
4027 when PCRE is compiled, and changed at run time. A patch from Google adds
4028 this functionality to the C++ interface.
4030 18. Changes to the handling of Unicode character properties:
4032 (a) Updated the table to Unicode 4.1.0.
4034 (b) Recognize characters that are not in the table as "Cn" (undefined).
4036 (c) I revised the way the table is implemented to a much improved format
4037 which includes recognition of ranges. It now supports the ranges that
4038 are defined in UnicodeData.txt, and it also amalgamates other
4039 characters into ranges. This has reduced the number of entries in the
4040 table from around 16,000 to around 3,000, thus reducing its size
4041 considerably. I realized I did not need to use a tree structure after
4042 all - a binary chop search is just as efficient. Having reduced the
4043 number of entries, I extended their size from 6 bytes to 8 bytes to
4044 allow for more data.
4046 (d) Added support for Unicode script names via properties such as \p{Han}.
4048 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4049 matching that character.
4051 20. When matching a repeated Unicode property with a minimum greater than zero,
4052 (for example \pL{2,}), PCRE could look past the end of the subject if it
4053 reached it while seeking the minimum number of characters. This could
4054 happen only if some of the characters were more than one byte long, because
4055 there is a check for at least the minimum number of bytes.
4057 21. Refactored the implementation of \p and \P so as to be more general, to
4058 allow for more different types of property in future. This has changed the
4059 compiled form incompatibly. Anybody with saved compiled patterns that use
4060 \p or \P will have to recompile them.
4062 22. Added "Any" and "L&" to the supported property types.
4064 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4065 but give a compile time error if the value is greater than 0xff.
4067 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4068 accidentally not being installed or uninstalled.
4070 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4071 made were to insert the current release number. This seemed silly, because
4072 it made things harder for people building PCRE on systems that don't run
4073 "configure". I have turned pcre.h into a distributed file, no longer built
4074 by "configure", with the version identification directly included. There is
4075 no longer a pcre.h.in file.
4077 However, this change necessitated a change to the pcre-config script as
4078 well. It is built from pcre-config.in, and one of the substitutions was the
4079 release number. I have updated configure.ac so that ./configure now finds
4080 the release number by grepping pcre.h.
4082 26. Added the ability to run the tests under valgrind.
4085 Version 6.4 05-Sep-05
4086 ---------------------
4088 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4089 "--" to be printed when multiple files were scanned, even when none of the
4090 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4091 consider it to be a bug, and have restored the previous behaviour.
4093 2. A couple of code tidies to get rid of compiler warnings.
4095 3. The pcretest program used to cheat by referring to symbols in the library
4096 whose names begin with _pcre_. These are internal symbols that are not
4097 really supposed to be visible externally, and in some environments it is
4098 possible to suppress them. The cheating is now confined to including
4099 certain files from the library's source, which is a bit cleaner.
4101 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4102 file's purpose clearer.
4104 5. Reorganized pcre_ucp_findchar().
4107 Version 6.3 15-Aug-05
4108 ---------------------
4110 1. The file libpcre.pc.in did not have general read permission in the tarball.
4112 2. There were some problems when building without C++ support:
4114 (a) If C++ support was not built, "make install" and "make test" still
4115 tried to test it.
4117 (b) There were problems when the value of CXX was explicitly set. Some
4118 changes have been made to try to fix these, and ...
4120 (c) --disable-cpp can now be used to explicitly disable C++ support.
4122 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4123 backslash in a target when C++ was disabled. This confuses some
4124 versions of "make", apparently. Using an intermediate variable solves
4125 this. (Same for CPP_LOBJ.)
4127 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4128 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4129 necessary on certain architectures.
4131 4. Added a setting of -export-symbols-regex to the link command to remove
4132 those symbols that are exported in the C sense, but actually are local
4133 within the library, and not documented. Their names all begin with
4134 "_pcre_". This is not a perfect job, because (a) we have to except some
4135 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4136 available (and never for static libraries). I have made a note to try to
4137 find a way round (a) in the future.
4140 Version 6.2 01-Aug-05
4141 ---------------------
4143 1. There was no test for integer overflow of quantifier values. A construction
4144 such as {1111111111111111} would give undefined results. What is worse, if
4145 a minimum quantifier for a parenthesized subpattern overflowed and became
4146 negative, the calculation of the memory size went wrong. This could have
4147 led to memory overwriting.
4149 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4151 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4152 operating environments where this matters.
4154 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4155 PCRE options from within the C++ wrapper.
4157 5. Named capturing subpatterns were not being correctly counted when a pattern
4158 was compiled. This caused two problems: (a) If there were more than 100
4159 such subpatterns, the calculation of the memory needed for the whole
4160 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4161 back references of the form \12, where the number was greater than 9, were
4162 not recognized as back references, even though there were sufficient
4163 previous subpatterns.
4165 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4166 versions of gcc, e.g. 2.95.4.
4169 Version 6.1 21-Jun-05
4170 ---------------------
4172 1. There was one reference to the variable "posix" in pcretest.c that was not
4173 surrounded by "#if !defined NOPOSIX".
4175 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4176 the cross-check on the old pcre_info() function, for the benefit of the
4177 cut-down version of PCRE that is currently imported into Exim.
4179 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4180 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4181 patterns that start (?i) but I don't think that matters. The use of (?i) is
4182 just an example; this all applies to the other options as well.
4184 4. Since libtool seems to echo the compile commands it is issuing, the output
4185 from "make" can be reduced a bit by putting "@" in front of each libtool
4186 compile command.
4188 5. Patch from the folks at Google for configure.in to be a bit more thorough
4189 in checking for a suitable C++ installation before trying to compile the
4190 C++ stuff. This should fix a reported problem when a compiler was present,
4191 but no suitable headers.
4193 6. The man pages all had just "PCRE" as their title. I have changed them to
4194 be the relevant file name. I have also arranged that these names are
4195 retained in the file doc/pcre.txt, which is a concatenation in text format
4196 of all the man pages except the little individual ones for each function.
4198 7. The NON-UNIX-USE file had not been updated for the different set of source
4199 files that come with release 6. I also added a few comments about the C++
4200 wrapper.
4203 Version 6.0 07-Jun-05
4204 ---------------------
4206 1. Some minor internal re-organization to help with my DFA experiments.
4208 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4209 didn't matter for the library itself when fully configured, but did matter
4210 when compiling without UCP support, or within Exim, where the ucp files are
4211 not imported.
4213 3. Refactoring of the library code to split up the various functions into
4214 different source modules. The addition of the new DFA matching code (see
4215 below) to a single monolithic source would have made it really too
4216 unwieldy, quite apart from causing all the code to be include in a
4217 statically linked application, when only some functions are used. This is
4218 relevant even without the DFA addition now that patterns can be compiled in
4219 one application and matched in another.
4221 The downside of splitting up is that there have to be some external
4222 functions and data tables that are used internally in different modules of
4223 the library but which are not part of the API. These have all had their
4224 names changed to start with "_pcre_" so that they are unlikely to clash
4225 with other external names.
4227 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4228 a different (DFA) algorithm. Although it is slower than the original
4229 function, it does have some advantages for certain types of matching
4230 problem.
4232 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4233 including restarting after a partial match.
4235 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4236 defined when compiling for Windows was sent to me. I have put it into the
4237 code, though I have no means of testing or verifying it.
4239 7. Added the pcre_refcount() auxiliary function.
4241 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4242 match before or at the first newline in the subject string. In pcretest,
4243 the /f option on a pattern can be used to set this.
4245 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4246 would behave wrongly. This has been present in PCRE since release 4.0.
4248 10. A number of changes to the pcregrep command:
4250 (a) Refactored how -x works; insert ^(...)$ instead of setting
4251 PCRE_ANCHORED and checking the length, in preparation for adding
4252 something similar for -w.
4254 (b) Added the -w (match as a word) option.
4256 (c) Refactored the way lines are read and buffered so as to have more
4257 than one at a time available.
4259 (d) Implemented a pcregrep test script.
4261 (e) Added the -M (multiline match) option. This allows patterns to match
4262 over several lines of the subject. The buffering ensures that at least
4263 8K, or the rest of the document (whichever is the shorter) is available
4264 for matching (and similarly the previous 8K for lookbehind assertions).
4266 (f) Changed the --help output so that it now says
4268 -w, --word-regex(p)
4270 instead of two lines, one with "regex" and the other with "regexp"
4271 because that confused at least one person since the short forms are the
4272 same. (This required a bit of code, as the output is generated
4273 automatically from a table. It wasn't just a text change.)
4275 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4276 option but starts with a hyphen. Could be a pattern or a path name
4277 starting with a hyphen, for instance.
4279 (h) "-" can be given as a file name to represent stdin.
4281 (i) When file names are being printed, "(standard input)" is used for
4282 the standard input, for compatibility with GNU grep. Previously
4283 "<stdin>" was used.
4285 (j) The option --label=xxx can be used to supply a name to be used for
4286 stdin when file names are being printed. There is no short form.
4288 (k) Re-factored the options decoding logic because we are going to add
4289 two more options that take data. Such options can now be given in four
4290 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4292 (l) Added the -A, -B, and -C options for requesting that lines of context
4293 around matches be printed.
4295 (m) Added the -L option to print the names of files that do not contain
4296 any matching lines, that is, the complement of -l.
4298 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4299 continue to scan other files.
4301 (o) The -s option was incorrectly implemented. For compatibility with other
4302 greps, it now suppresses the error message for a non-existent or non-
4303 accessible file (but not the return code). There is a new option called
4304 -q that suppresses the output of matching lines, which was what -s was
4305 previously doing.
4307 (p) Added --include and --exclude options to specify files for inclusion
4308 and exclusion when recursing.
4310 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4311 Hopefully, it now does.
4313 12. Missing cast in pcre_study().
4315 13. Added an "uninstall" target to the makefile.
4317 14. Replaced "extern" in the function prototypes in Makefile.in with
4318 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4319 world, but is set differently for Windows.
4321 15. Added a second compiling function called pcre_compile2(). The only
4322 difference is that it has an extra argument, which is a pointer to an
4323 integer error code. When there is a compile-time failure, this is set
4324 non-zero, in addition to the error test pointer being set to point to an
4325 error message. The new argument may be NULL if no error number is required
4326 (but then you may as well call pcre_compile(), which is now just a
4327 wrapper). This facility is provided because some applications need a
4328 numeric error indication, but it has also enabled me to tidy up the way
4329 compile-time errors are handled in the POSIX wrapper.
4331 16. Added VPATH=.libs to the makefile; this should help when building with one
4332 prefix path and installing with another. (Or so I'm told by someone who
4333 knows more about this stuff than I do.)
4335 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4336 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4337 match everything, including newlines. This is not POSIX-compatible, but
4338 somebody wanted the feature. From pcretest it can be activated by using
4339 both the P and the s flags.
4341 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4343 19. libpcre.pc was being incorrectly installed as executable.
4345 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4346 it now also looks for '\r' so that it will work unmodified on Windows.
4348 21. Added Google's contributed C++ wrapper to the distribution.
4350 22. Added some untidy missing memory free() calls in pcretest, to keep
4351 Electric Fence happy when testing.
4355 Version 5.0 13-Sep-04
4356 ---------------------
4358 1. Internal change: literal characters are no longer packed up into items
4359 containing multiple characters in a single byte-string. Each character
4360 is now matched using a separate opcode. However, there may be more than one
4361 byte in the character in UTF-8 mode.
4363 2. The pcre_callout_block structure has two new fields: pattern_position and
4364 next_item_length. These contain the offset in the pattern to the next match
4365 item, and its length, respectively.
4367 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4368 insertion of callouts before each pattern item. Added the /C option to
4369 pcretest to make use of this.
4371 4. On the advice of a Windows user, the lines
4373 #if defined(_WIN32) || defined(WIN32)
4374 _setmode( _fileno( stdout ), 0x8000 );
4375 #endif /* defined(_WIN32) || defined(WIN32) */
4377 have been added to the source of pcretest. This apparently does useful
4378 magic in relation to line terminators.
4380 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4381 for the benefit of those environments where the "b" makes a difference.
4383 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4384 to know about it. I have put a hack into configure.in that adds in code
4385 to set GCC=yes if CC=icc. This seems to end up at a point in the
4386 generated configure script that is early enough to affect the setting of
4387 compiler options, which is what is needed, but I have no means of testing
4388 whether it really works. (The user who reported this had patched the
4389 generated configure script, which of course I cannot do.)
4391 LATER: After change 22 below (new libtool files), the configure script
4392 seems to know about icc (and also ecc). Therefore, I have commented out
4393 this hack in configure.in.
4395 7. Added support for pkg-config (2 patches were sent in).
4397 8. Negated POSIX character classes that used a combination of internal tables
4398 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4399 [[:^ascii]]. Typically, they would match almost any characters. The other
4400 POSIX classes were not broken in this way.
4402 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4403 to find the match, as PCRE was deluded into thinking that the match had to
4404 start at the start point or following a newline. The same bug applied to
4405 patterns with negative forward assertions or any backward assertions
4406 preceding ".*" at the start, unless the pattern required a fixed first
4407 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4409 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4410 starting at the last subject character, bytes beyond the end of the subject
4411 string were read.
4413 11. Renamed the variable "class" as "classbits" to make life easier for C++
4414 users. (Previously there was a macro definition, but it apparently wasn't
4415 enough.)
4417 12. Added the new field "tables" to the extra data so that tables can be passed
4418 in at exec time, or the internal tables can be re-selected. This allows
4419 a compiled regex to be saved and re-used at a later time by a different
4420 program that might have everything at different addresses.
4422 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4423 -R library as well as a -L library.
4425 14. The debugging options of pcretest (-d on the command line or D on a
4426 pattern) showed incorrect output for anything following an extended class
4427 that contained multibyte characters and which was followed by a quantifier.
4429 15. Added optional support for general category Unicode character properties
4430 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4431 support. It adds about 90K to the size of the library. The meanings of the
4432 inbuilt class escapes such as \d and \s have NOT been changed.
4434 16. Updated pcredemo.c to include calls to free() to release the memory for the
4435 compiled pattern.
4437 17. The generated file chartables.c was being created in the source directory
4438 instead of in the building directory. This caused the build to fail if the
4439 source directory was different from the building directory, and was
4440 read-only.
4442 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4443 file. No doubt somebody will tell me if they don't make sense... Also added
4444 Dan Mooney's comments about building on OpenVMS.
4446 19. Added support for partial matching via the PCRE_PARTIAL option for
4447 pcre_exec() and the \P data escape in pcretest.
4449 20. Extended pcretest with 3 new pattern features:
4451 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4452 write the compiled pattern to the file whose name is "rest-of-line".
4453 This is a straight binary dump of the data, with the saved pointer to
4454 the character tables forced to be NULL. The study data, if any, is
4455 written too. After writing, pcretest reads a new pattern.
4457 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4458 compiled pattern from the given file. There must not be any
4459 occurrences of "<" in the file name (pretty unlikely); if there are,
4460 pcretest will instead treat the initial "<" as a pattern delimiter.
4461 After reading in the pattern, pcretest goes on to read data lines as
4462 usual.
4464 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4465 and 16-bit fields in a compiled pattern, to simulate a pattern that
4466 was compiled on a host of opposite endianness.
4468 21. The pcre-exec() function can now cope with patterns that were compiled on
4469 hosts of opposite endianness, with this restriction:
4471 As for any compiled expression that is saved and used later, the tables
4472 pointer field cannot be preserved; the extra_data field in the arguments
4473 to pcre_exec() should be used to pass in a tables address if a value
4474 other than the default internal tables were used at compile time.
4476 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4477 now diagnosed as an error. Previously, most of the time, a negative number
4478 would have been treated as zero, but if in addition "ovector" was passed as
4479 NULL, a crash could occur.
4481 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4482 new versions from the libtool 1.5 distribution (the last one is a copy of
4483 a file called libtool.m4). This seems to have fixed the need to patch
4484 "configure" to support Darwin 1.3 (which I used to do). However, I still
4485 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4486 workstation).
4488 24. Changed the PCRE licence to be the more standard "BSD" licence.
4491 Version 4.5 01-Dec-03
4492 ---------------------
4494 1. There has been some re-arrangement of the code for the match() function so
4495 that it can be compiled in a version that does not call itself recursively.
4496 Instead, it keeps those local variables that need separate instances for
4497 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4498 needs to "recurse". Keeping track of where control must go is done by means
4499 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4500 hide most of the details from the main code, and operates only if
4501 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4502 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4503 operating.
4505 To make it easier for callers to provide specially tailored get/free
4506 functions for this usage, two new functions, pcre_stack_malloc, and
4507 pcre_stack_free, are used. They are always called in strict stacking order,
4508 and the size of block requested is always the same.
4510 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4511 PCRE has been compiled to use the stack or the heap for recursion. The
4512 -C option of pcretest uses this to show which version is compiled.
4514 A new data escape \S, is added to pcretest; it causes the amounts of store
4515 obtained and freed by both kinds of malloc/free at match time to be added
4516 to the output.
4518 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4519 what's available on my current Linux desktop machine.
4521 3. When matching a UTF-8 string, the test for a valid string at the start has
4522 been extended. If start_offset is not zero, PCRE now checks that it points
4523 to a byte that is the start of a UTF-8 character. If not, it returns
4524 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4525 this is necessary because there may be backward assertions in the pattern.
4526 When matching the same subject several times, it may save resources to use
4527 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4529 4. The code for checking the validity of UTF-8 strings has been tightened so
4530 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4531 containing "overlong sequences".
4533 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4534 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4535 should have been "&", but it just so happened that all the cases this let
4536 through by mistake were picked up later in the function.
4538 6. I had used a variable called "isblank" - this is a C99 function, causing
4539 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4541 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4542 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4543 so the version is shown; (c) stop "make test" echoing "./RunTest".
4545 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4547 9. The prototype for memmove() for systems that don't have it was using
4548 size_t, but the inclusion of the header that defines size_t was later. I've
4549 moved the #includes for the C headers earlier to avoid this.
4551 10. Added some adjustments to the code to make it easier to compiler on certain
4552 special systems:
4554 (a) Some "const" qualifiers were missing.
4555 (b) Added the macro EXPORT before all exported functions; by default this
4556 is defined to be empty.
4557 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4558 that it reads its output file name as an argument instead of writing
4559 to the standard output and assuming this can be redirected.
4561 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4562 class containing characters with values greater than 255, PCRE compilation
4563 went into a loop.
4565 12. A recursive reference to a subpattern that was within another subpattern
4566 that had a minimum quantifier of zero caused PCRE to crash. For example,
4567 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4568 recursion. If the recursively-called subpattern itself had a zero repeat,
4569 that was OK.
4571 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4572 buffer into which it was copied (for escape processing) was still set at
4573 1024, so long lines caused crashes.
4575 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4576 "internal error: code overflow...". This applied to any character class
4577 that was followed by a possessive quantifier.
4579 15. Modified the Makefile to add libpcre.la as a prerequisite for
4580 libpcreposix.la because I was told this is needed for a parallel build to
4581 work.
4583 16. If a pattern that contained .* following optional items at the start was
4584 studied, the wrong optimizing data was generated, leading to matching
4585 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4586 matching string must start with a or b or c. The correct conclusion for
4587 this pattern is that a match can start with any character.
4590 Version 4.4 13-Aug-03
4591 ---------------------
4593 1. In UTF-8 mode, a character class containing characters with values between
4594 127 and 255 was not handled correctly if the compiled pattern was studied.
4595 In fixing this, I have also improved the studying algorithm for such
4596 classes (slightly).
4598 2. Three internal functions had redundant arguments passed to them. Removal
4599 might give a very teeny performance improvement.
4601 3. Documentation bug: the value of the capture_top field in a callout is *one
4602 more than* the number of the hightest numbered captured substring.
4604 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4605 in incorrectly linking with a previously installed version. They now link
4606 explicitly with libpcre.la.
4608 5. configure.in no longer needs to recognize Cygwin specially.
4610 6. A problem in pcre.in for Windows platforms is fixed.
4612 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4613 pcretest, it used to include the size of the study block as part of its
4614 output. Unfortunately, the structure contains a field that has a different
4615 size on different hardware architectures. This meant that the tests that
4616 showed this size failed. As the block is currently always of a fixed size,
4617 this information isn't actually particularly useful in pcretest output, so
4618 I have just removed it.
4620 8. Three pre-processor statements accidentally did not start in column 1.
4621 Sadly, there are *still* compilers around that complain, even though
4622 standard C has not required this for well over a decade. Sigh.
4624 9. In pcretest, the code for checking callouts passed small integers in the
4625 callout_data field, which is a void * field. However, some picky compilers
4626 complained about the casts involved for this on 64-bit systems. Now
4627 pcretest passes the address of the small integer instead, which should get
4628 rid of the warnings.
4630 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4631 both compile and run time, and gives an error if an invalid UTF-8 sequence
4632 is found. There is a option for disabling this check in cases where the
4633 string is known to be correct and/or the maximum performance is wanted.
4635 11. In response to a bug report, I changed one line in Makefile.in from
4637 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4638 to
4639 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4641 to look similar to other lines, but I have no way of telling whether this
4642 is the right thing to do, as I do not use Windows. No doubt I'll get told
4643 if it's wrong...
4646 Version 4.3 21-May-03
4647 ---------------------
4649 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4650 Makefile.
4652 2. Some refactoring to improve the quality of the code:
4654 (i) The utf8_table... variables are now declared "const".
4656 (ii) The code for \cx, which used the "case flipping" table to upper case
4657 lower case letters, now just substracts 32. This is ASCII-specific,
4658 but the whole concept of \cx is ASCII-specific, so it seems
4659 reasonable.
4661 (iii) PCRE was using its character types table to recognize decimal and
4662 hexadecimal digits in the pattern. This is silly, because it handles
4663 only 0-9, a-f, and A-F, but the character types table is locale-
4664 specific, which means strange things might happen. A private
4665 table is now used for this - though it costs 256 bytes, a table is
4666 much faster than multiple explicit tests. Of course, the standard
4667 character types table is still used for matching digits in subject
4668 strings against \d.
4670 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4671 ending in _t are). So I've renamed it as ESC_tee.
4673 3. The first argument for regexec() in the POSIX wrapper should have been
4674 defined as "const".
4676 4. Changed pcretest to use malloc() for its buffers so that they can be
4677 Electric Fenced for debugging.
4679 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4680 to read one or more bytes before the start of the subject string. Often this
4681 had no effect on PCRE's behaviour, but in some circumstances it could
4682 provoke a segmentation fault.
4684 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4685 to try to read one or more bytes before the start of the subject string.
4687 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4688 UTF-8 support could misbehave in various ways if the subject string
4689 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4690 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4691 back over UTF-8 characters.)
4694 Version 4.2 14-Apr-03
4695 ---------------------
4697 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4699 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4700 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4701 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4702 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4703 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4705 Note: automatic setting of the BUILD variables is not yet working
4706 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4707 compile-time but not at link-time
4708 [LINK]: use for linking executables only
4709 make different versions for Windows and non-Windows
4710 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4711 libraries
4712 [LINK_FOR_BUILD]: new variable
4713 [OBJEXT]: use throughout
4714 [EXEEXT]: use throughout
4715 <winshared>: new target
4716 <wininstall>: new target
4717 <dftables.o>: use native compiler
4718 <dftables>: use native linker
4719 <install>: handle Windows platform correctly
4720 <clean>: ditto
4721 <check>: ditto
4722 copy DLL to top builddir before testing
4724 As part of these changes, -no-undefined was removed again. This was reported
4725 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4726 in any case.
4728 3. Some tidies to get rid of compiler warnings:
4730 . In the match_data structure, match_limit was an unsigned long int, whereas
4731 match_call_count was an int. I've made them both unsigned long ints.
4733 . In pcretest the fact that a const uschar * doesn't automatically cast to
4734 a void * provoked a warning.
4736 . Turning on some more compiler warnings threw up some "shadow" variables
4737 and a few more missing casts.
4739 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4740 option, a class that contained a single character with a value between 128
4741 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4743 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4744 option, a class that contained several characters, but with at least one
4745 whose value was between 128 and 255 caused PCRE to crash.
4748 Version 4.1 12-Mar-03
4749 ---------------------
4751 1. Compiling with gcc -pedantic found a couple of places where casts were
4752 needed, and a string in dftables.c that was longer than standard compilers are
4753 required to support.
4755 2. Compiling with Sun's compiler found a few more places where the code could
4756 be tidied up in order to avoid warnings.
4758 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4759 first of these names is deprecated in the latest Autoconf in favour of the name
4760 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4761 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4762 analogy I have changed it to CFLAGS_FOR_BUILD.
4764 4. Added -no-undefined to the linking command in the Makefile, because this is
4765 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4766 linking step for the pcreposix library.
4768 5. PCRE was failing to diagnose the case of two named groups with the same
4769 name.
4771 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4772 literal character that is needed in the subject for a match, and scans along to
4773 ensure that it is present before embarking on the full matching process. This
4774 saves time in cases of nested unlimited repeats that are never going to match.
4775 Problem: the scan can take a lot of time if the subject is very long (e.g.
4776 megabytes), thus penalizing straightforward matches. It is now done only if the
4777 amount of subject to be scanned is less than 1000 bytes.
4779 7. A lesser problem with the same optimization is that it was recording the
4780 first character of an anchored pattern as "needed", thus provoking a search
4781 right along the subject, even when the first match of the pattern was going to
4782 fail. The "needed" character is now not set for anchored patterns, unless it
4783 follows something in the pattern that is of non-fixed length. Thus, it still
4784 fulfils its original purpose of finding quick non-matches in cases of nested
4785 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4788 Version 4.0 17-Feb-03
4789 ---------------------
4791 1. If a comment in an extended regex that started immediately after a meta-item
4792 extended to the end of string, PCRE compiled incorrect data. This could lead to
4793 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4795 2. Moved to autoconf 2.53 and libtool 1.4.2.
4797 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4798 the special perltest8 script is no longer needed - all the tests can be run
4799 from a single perltest script.
4801 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4802 by \s. It has now been removed in PCRE. This means it isn't recognized as
4803 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4804 class [:space:] *does* include VT, thereby creating a mess.
4806 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4807 space and tab.
4809 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4810 its new features into the main test script, reducing the number of scripts.
4812 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4813 were backward compatible, and made the (?i) apply to the whole pattern, as if
4814 /i were given. Now it behaves more logically, and applies the option setting
4815 only to what follows. PCRE has been changed to follow suit. However, if it
4816 finds options settings right at the start of the pattern, it extracts them into
4817 the global options, as before. Thus, they show up in the info data.
4819 8. Added support for the \Q...\E escape sequence. Characters in between are
4820 treated as literals. This is slightly different from Perl in that $ and @ are
4821 also handled as literals inside the quotes. In Perl, they will cause variable
4822 interpolation. Note the following examples:
4824 Pattern PCRE matches Perl matches
4826 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4827 \Qabc\$xyz\E abc\$xyz abc\$xyz
4828 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4830 For compatibility with Perl, \Q...\E sequences are recognized inside character
4831 classes as well as outside them.
4833 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4834 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4835 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4836 signed/unsigned warnings.
4838 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4839 option for pcretest, so I've replaced it by a simple function that does just
4840 that job.
4842 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4843 "pcregrep -".
4845 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4846 Java package. This provides some syntactic sugar for simple cases of what my
4847 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4848 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4849 item, you can use this simplified notation. Note that only makes sense with
4850 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4851 greediness, whatever the setting of the PCRE_UNGREEDY option.
4853 13. A change of greediness default within a pattern was not taking effect at
4854 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4855 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4856 was abstracted outside.
4858 14. PCRE now supports the \G assertion. It is true when the current matching
4859 position is at the start point of the match. This differs from \A when the
4860 starting offset is non-zero. Used with the /g option of pcretest (or similar
4861 code), it works in the same way as it does for Perl's /g option. If all
4862 alternatives of a regex begin with \G, the expression is anchored to the start
4863 match position, and the "anchored" flag is set in the compiled expression.
4865 15. Some bugs concerning the handling of certain option changes within patterns
4866 have been fixed. These applied to options other than (?ims). For example,
4867 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4868 been the other way round. Some of this was related to change 7 above.
4870 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4871 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4872 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4873 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4875 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4876 mode. Unlike ".", it always matches newline, whatever the setting of
4877 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4878 assertions. Perl allows it, but it doesn't (in general) work because it can't
4879 calculate the length of the lookbehind. At least, that's the case for Perl
4880 5.8.0 - I've been told they are going to document that it doesn't work in
4881 future.
4883 18. Added an error diagnosis for escapes that PCRE does not support: these are
4884 \L, \l, \N, \P, \p, \U, \u, and \X.
4886 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4887 reading past the end of the pattern in cases such as /[abcd/.
4889 20. PCRE was getting more memory than necessary for patterns with classes that
4890 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4892 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4893 compiling PCRE for use with Virtual Pascal.
4895 22. Small fix to the Makefile to make it work properly if the build is done
4896 outside the source tree.
4898 23. Added a new extension: a condition to go with recursion. If a conditional
4899 subpattern starts with (?(R) the "true" branch is used if recursion has
4900 happened, whereas the "false" branch is used only at the top level.
4902 24. When there was a very long string of literal characters (over 255 bytes
4903 without UTF support, over 250 bytes with UTF support), the computation of how
4904 much memory was required could be incorrect, leading to segfaults or other
4905 strange effects.
4907 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4908 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4909 there was a subsequent back reference to those brackets. This meant that, for
4910 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4911 possible to check for precisely this case. All we can do is abandon the
4912 optimization if .* occurs inside capturing brackets when there are any back
4913 references whatsoever. (See below for a better fix that came later.)
4915 26. The handling of the optimization for finding the first character of a
4916 non-anchored pattern, and for finding a character that is required later in the
4917 match were failing in some cases. This didn't break the matching; it just
4918 failed to optimize when it could. The way this is done has been re-implemented.
4920 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4922 28. Added a new feature that provides some of the functionality that Perl
4923 provides with (?{...}). The facility is termed a "callout". The way it is done
4924 in PCRE is for the caller to provide an optional function, by setting
4925 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4926 global variable. By default it is unset, which disables all calling out. To get
4927 the function called, the regex must include (?C) at appropriate points. This
4928 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4929 This provides a means of identifying different callout points. When PCRE
4930 reaches such a point in the regex, if pcre_callout has been set, the external
4931 function is called. It is provided with data in a structure called
4932 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4933 matching continues; if it returns a non-zero value, the match at the current
4934 point fails. However, backtracking will occur if possible. [This was changed
4935 later and other features added - see item 49 below.]
4937 29. pcretest is upgraded to test the callout functionality. It provides a
4938 callout function that displays information. By default, it shows the start of
4939 the match and the current position in the text. There are some new data escapes
4940 to vary what happens:
4942 \C+ in addition, show current contents of captured substrings
4943 \C- do not supply a callout function
4944 \C!n return 1 when callout number n is reached
4945 \C!n!m return 1 when callout number n is reached for the mth time
4947 30. If pcregrep was called with the -l option and just a single file name, it
4948 output "<stdin>" if a match was found, instead of the file name.
4950 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4951 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4952 pcre_exec(). This saves a malloc/free per call. The default value of
4953 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4954 when configuring.
4956 32. The default maximum size of a compiled pattern is 64K. There have been a
4957 few cases of people hitting this limit. The code now uses macros to handle the
4958 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4959 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4960 configuring. Tests 2 and 5 work only with 2-byte links because they output
4961 debugging information about compiled patterns.
4963 33. Internal code re-arrangements:
4965 (a) Moved the debugging function for printing out a compiled regex into
4966 its own source file (printint.c) and used #include to pull it into
4967 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4968 separate copies.
4970 (b) Defined the list of op-code names for debugging as a macro in
4971 internal.h so that it is next to the definition of the opcodes.
4973 (c) Defined a table of op-code lengths for simpler skipping along compiled
4974 code. This is again a macro in internal.h so that it is next to the
4975 definition of the opcodes.
4977 34. Added support for recursive calls to individual subpatterns, along the
4978 lines of Robin Houston's patch (but implemented somewhat differently).
4980 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4981 allow it to read and process whole directories in Win32. This code was
4982 contributed by Lionel Fourquaux; it has not been tested by me.
4984 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4985 used to name a group. Names consist of alphanumerics and underscores, and must
4986 be unique. Back references use the syntax (?P=name) and recursive calls use
4987 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4988 numbers. The function pcre_fullinfo() can be used after compilation to extract
4989 a name/number map. There are three relevant calls:
4991 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4992 PCRE_INFO_NAMECOUNT yields the number of entries
4993 PCRE_INFO_NAMETABLE yields a pointer to the map.
4995 The map is a vector of fixed-size entries. The size of each entry depends on
4996 the length of the longest name used. The first two bytes of each entry are the
4997 group number, most significant byte first. There follows the corresponding
4998 name, zero terminated. The names are in alphabetical order.
5000 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
5001 case instead of 255. Making it the same both with and without UTF-8 support
5002 means that the same test output works with both.
5004 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
5005 calling malloc() with a zero argument.
5007 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
5008 optimization. I've improved things by keeping a bitmap of backreferences with
5009 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
5010 fact referenced, the optimization can be applied. It is unlikely that a
5011 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
5012 the match to follow \n) will appear inside brackets with a number greater than
5013 31, but if it does, any back reference > 31 suppresses the optimization.
5015 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
5016 of disabling numbered capturing parentheses. Any opening parenthesis that is
5017 not followed by ? behaves as if it were followed by ?: but named parentheses
5018 can still be used for capturing (and they will acquire numbers in the usual
5019 way).
5021 41. Redesigned the return codes from the match() function into yes/no/error so
5022 that errors can be passed back from deep inside the nested calls. A malloc
5023 failure while inside a recursive subpattern call now causes the
5024 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
5026 42. It is now possible to set a limit on the number of times the match()
5027 function is called in a call to pcre_exec(). This facility makes it possible to
5028 limit the amount of recursion and backtracking, though not in a directly
5029 obvious way, because the match() function is used in a number of different
5030 circumstances. The count starts from zero for each position in the subject
5031 string (for non-anchored patterns). The default limit is, for compatibility, a
5032 large number, namely 10 000 000. You can change this in two ways:
5034 (a) When configuring PCRE before making, you can use --with-match-limit=n
5035 to set a default value for the compiled library.
5037 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5038 a different value is set. See 45 below.
5040 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5042 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5043 of things that can be changed at compile time. The first argument specifies
5044 what is wanted and the second points to where the information is to be placed.
5045 The current list of available information is:
5049 The output is an integer that is set to one if UTF-8 support is available;
5050 otherwise it is set to zero.
5054 The output is an integer that it set to the value of the code that is used for
5055 newline. It is either LF (10) or CR (13).
5059 The output is an integer that contains the number of bytes used for internal
5060 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5064 The output is an integer that contains the threshold above which the POSIX
5065 interface uses malloc() for output vectors. See item 31 above.
5069 The output is an unsigned integer that contains the default limit of the number
5070 of match() calls in a pcre_exec() execution. See 42 above.
5072 44. pcretest has been upgraded by the addition of the -C option. This causes it
5073 to extract all the available output from the new pcre_config() function, and to
5074 output it. The program then exits immediately.
5076 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5077 order to support additional features. One way would have been to define
5078 pcre_exec2() (for example) with extra arguments, but this would not have been
5079 extensible, and would also have required all calls to the original function to
5080 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5081 is used for passing in "extra" data from pcre_study().
5083 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5084 contains the following fields:
5086 flags a bitmap indicating which of the following fields are set
5087 study_data opaque data from pcre_study()
5088 match_limit a way of specifying a limit on match() calls for a specific
5089 call to pcre_exec()
5090 callout_data data for callouts (see 49 below)
5092 The flag bits are also defined in pcre.h, and are
5098 The pcre_study() function now returns one of these new pcre_extra blocks, with
5099 the actual study data pointed to by the study_data field, and the
5100 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5101 before. That is, this change is entirely upwards-compatible and requires no
5102 change to existing code.
5104 If you want to pass in additional data to pcre_exec(), you can either place it
5105 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5106 block.
5108 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5109 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5110 times with different match limits, until it finds the minimum value needed for
5111 pcre_exec() to complete. The value is then output. This can be instructive; for
5112 most simple matches the number is quite small, but for pathological cases it
5113 gets very large very quickly.
5115 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5116 returns the size of the data block pointed to by the study_data field in a
5117 pcre_extra block, that is, the value that was passed as the argument to
5118 pcre_malloc() when PCRE was getting memory in which to place the information
5119 created by pcre_study(). The fourth argument should point to a size_t variable.
5120 pcretest has been extended so that this information is shown after a successful
5121 pcre_study() call when information about the compiled regex is being displayed.
5123 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5124 because what follows is always an absolute path. (Later: it turns out that this
5125 is more than cosmetic for MinGW, because it doesn't like empty path
5126 components.)
5128 49. Some changes have been made to the callout feature (see 28 above):
5130 (i) A callout function now has three choices for what it returns:
5132 0 => success, carry on matching
5133 > 0 => failure at this point, but backtrack if possible
5134 < 0 => serious error, return this value from pcre_exec()
5136 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5137 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5138 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5139 use by callout functions. It will never be used by PCRE itself.
5141 (ii) The pcre_extra structure (see 45 above) has a void * field called
5142 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5143 pcre_callout_block structure has a field of the same name. The contents of
5144 the field passed in the pcre_extra structure are passed to the callout
5145 function in the corresponding field in the callout block. This makes it
5146 easier to use the same callout-containing regex from multiple threads. For
5147 testing, the pcretest program has a new data escape
5149 \C*n pass the number n (may be negative) as callout_data
5151 If the callout function in pcretest receives a non-zero value as
5152 callout_data, it returns that value.
5154 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5155 there were some redundant $(CFLAGS) in commands that are now specified as
5156 $(LINK), which already includes $(CFLAGS).
5158 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5159 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5160 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5161 one-byte characters throughout. Note that case-insensitive matching applies
5162 only to characters whose values are less than 256. PCRE doesn't support the
5163 notion of cases for higher-valued characters.
5165 (i) A character class whose characters are all within 0-255 is handled as
5166 a bit map, and the map is inverted for negative classes. Previously, a
5167 character > 255 always failed to match such a class; however it should
5168 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5170 (ii) A negated character class with a single character < 255 is coded as
5171 "not this character" (OP_NOT). This wasn't working properly when the test
5172 character was multibyte, either singly or repeated.
5174 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5175 mode, for example: \x{100}{2,3}.
5177 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5178 singly or repeated) now correctly test multibyte characters. However,
5179 PCRE doesn't recognize any characters with values greater than 255 as
5180 digits, spaces, or word characters. Such characters always match \D, \S,
5181 and \W, and never match \d, \s, or \w.
5183 (v) Classes may now contain characters and character ranges with values
5184 greater than 255. For example: [ab\x{100}-\x{400}].
5186 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5187 PCRE in UTF-8 mode.
5189 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5190 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5191 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5192 value.)
5194 53. The single man page has become too large. I have therefore split it up into
5195 a number of separate man pages. These also give rise to individual HTML pages;
5196 these are now put in a separate directory, and there is an index.html page that
5197 lists them all. Some hyperlinking between the pages has been installed.
5199 54. Added convenience functions for handling named capturing parentheses.
5201 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5202 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5203 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5204 are faulted.
5206 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5207 calling configure. These values are used when compiling the dftables.c program
5208 which is run to generate the source of the default character tables. They
5209 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5210 you will need to set these values.
5212 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5215 Version 3.9 02-Jan-02
5216 ---------------------
5218 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5220 2. If --disable-static was given, the building process failed when trying to
5221 build pcretest and pcregrep. (For some reason it was using libtool to compile
5222 them, which is not right, as they aren't part of the library.)
5225 Version 3.8 18-Dec-01
5226 ---------------------
5228 1. The experimental UTF-8 code was completely screwed up. It was packing the
5229 bytes in the wrong order. How dumb can you get?
5232 Version 3.7 29-Oct-01
5233 ---------------------
5235 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5236 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5237 this didn't happen under Solaris 8, where I normally test things.
5239 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5240 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5241 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5242 this fix an hour or so after the initial 3.7 release.)
5245 Version 3.6 23-Oct-01
5246 ---------------------
5248 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5249 offsets passed as NULL with zero offset count.
5251 2. The config.guess and config.sub files had not been updated when I moved to
5252 the latest autoconf.
5255 Version 3.5 15-Aug-01
5256 ---------------------
5258 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5259 had been forgotten.
5261 2. By using declared but undefined structures, we can avoid using "void"
5262 definitions in pcre.h while keeping the internal definitions of the structures
5263 private.
5265 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5266 user point of view, this means that both static and shared libraries are built
5267 by default, but this can be individually controlled. More of the work of
5268 handling this static/shared cases is now inside libtool instead of PCRE's make
5269 file.
5271 4. The pcretest utility is now installed along with pcregrep because it is
5272 useful for users (to test regexs) and by doing this, it automatically gets
5273 relinked by libtool. The documentation has been turned into a man page, so
5274 there are now .1, .txt, and .html versions in /doc.
5276 5. Upgrades to pcregrep:
5277 (i) Added long-form option names like gnu grep.
5278 (ii) Added --help to list all options with an explanatory phrase.
5279 (iii) Added -r, --recursive to recurse into sub-directories.
5280 (iv) Added -f, --file to read patterns from a file.
5282 6. pcre_exec() was referring to its "code" argument before testing that
5283 argument for NULL (and giving an error if it was NULL).
5285 7. Upgraded Makefile.in to allow for compiling in a different directory from
5286 the source directory.
5288 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5289 options bits, the pointer it was passed was to an int instead of to an unsigned
5290 long int. This mattered only on 64-bit systems.
5292 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5293 generated) instead of pcre.in, which it its source. Also made the same change
5294 in several of the .c files.
5296 10. A new release of gcc defines printf() as a macro, which broke pcretest
5297 because it had an ifdef in the middle of a string argument for printf(). Fixed
5298 by using separate calls to printf().
5300 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5301 script, to force use of CR or LF instead of \n in the source. On non-Unix
5302 systems, the value can be set in config.h.
5304 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5305 absolute limit. Changed the text of the error message to make this clear, and
5306 likewise updated the man page.
5308 13. The limit of 99 on the number of capturing subpatterns has been removed.
5309 The new limit is 65535, which I hope will not be a "real" limit.
5312 Version 3.4 22-Aug-00
5313 ---------------------
5315 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5317 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5320 Version 3.3 01-Aug-00
5321 ---------------------
5323 1. If an octal character was given, but the value was greater than \377, it
5324 was not getting masked to the least significant bits, as documented. This could
5325 lead to crashes in some systems.
5327 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5328 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5330 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5331 These just pass their arguments on to (pcre_free)(), but they are provided
5332 because some uses of PCRE bind it to non-C systems that can call its functions,
5333 but cannot call free() or pcre_free() directly.
5335 4. Add "make test" as a synonym for "make check". Corrected some comments in
5336 the Makefile.
5338 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5339 Makefile.
5341 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5342 command called pgrep for grepping around the active processes.
5344 7. Added the beginnings of support for UTF-8 character strings.
5346 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5347 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5348 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5349 out for the ar command.)
5352 Version 3.2 12-May-00
5353 ---------------------
5355 This is purely a bug fixing release.
5357 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5358 of ZA. This was just one example of several cases that could provoke this bug,
5359 which was introduced by change 9 of version 2.00. The code for breaking
5360 infinite loops after an iteration that matches an empty string was't working
5361 correctly.
5363 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5364 when matched against abbab (for example). After matching an empty string, it
5365 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5366 caused it to match further down the string than it should.
5368 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5369 was there because it doesn't seem to be needed, and it causes trouble on some
5370 systems, as it is not a Standard C header. It has been removed.
5372 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5373 were reported on the Macintosh. The changes were from
5375 while ((c = *(++ptr)) != 0 && c != '\n');
5376 to
5377 while ((c = *(++ptr)) != 0 && c != '\n') ;
5379 Totally extraordinary, but if that's what it takes...
5381 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5382 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5383 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5384 assumes the way PCRE uses memmove() (always moving upwards).
5386 6. PCRE is being used in one environment where strchr() is not available. There
5387 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5388 faster code anyway.
5391 Version 3.1 09-Feb-00
5392 ---------------------
5394 The only change in this release is the fixing of some bugs in Makefile.in for
5395 the "install" target:
5397 (1) It was failing to install pcreposix.h.
5399 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5402 Version 3.0 01-Feb-00
5403 ---------------------
5405 1. Add support for the /+ modifier to perltest (to output $` like it does in
5406 pcretest).
5408 2. Add support for the /g modifier to perltest.
5410 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5411 matches null strings.
5413 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5414 pattern. Perl treats empty patterns specially - it reuses the most recent
5415 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5416 effect.
5418 5. The POSIX interface was broken in that it was just handing over the POSIX
5419 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5420 required a bigger vector, with some working space on the end. This means that
5421 the POSIX wrapper now has to get and free some memory, and copy the results.
5423 6. Added some simple autoconf support, placing the test data and the
5424 documentation in separate directories, re-organizing some of the
5425 information files, and making it build pcre-config (a GNU standard). Also added
5426 libtool support for building PCRE as a shared library, which is now the
5427 default.
5429 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5430 09 are not valid octal constants. Single digits will be used for minor values
5431 less than 10.
5433 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5434 existing programs that set these in the POSIX interface can use PCRE without
5435 modification.
5437 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5438 return all that pcre_info() returns, plus additional data. The pcre_info()
5439 function is retained for compatibility, but is considered to be obsolete.
5441 10. Added experimental recursion feature (?R) to handle one common case that
5442 Perl 5.6 will be able to do with (?p{...}).
5444 11. Added support for POSIX character classes like [:alpha:], which Perl is
5445 adopting.
5448 Version 2.08 31-Aug-99
5449 ----------------------
5451 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5452 trying to match at the startoffset position, but instead was moving forward to
5453 the next newline as if a previous match had failed.
5455 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5456 and could get into a loop if a null string was matched other than at the start
5457 of the subject.
5459 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5460 be distinguished at compile time, and for completeness also added PCRE_DATE.
5462 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5463 in GnuWin32 environments.
5466 Version 2.07 29-Jul-99
5467 ----------------------
5469 1. The documentation is now supplied in plain text form and HTML as well as in
5470 the form of man page sources.
5472 2. C++ compilers don't like assigning (void *) values to other pointer types.
5473 In particular this affects malloc(). Although there is no problem in Standard
5474 C, I've put in casts to keep C++ compilers happy.
5476 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5477 should be (const char *).
5479 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5480 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5481 However, I haven't made this a standard facility. The documentation doesn't
5482 mention it, and the Makefile doesn't support it.
5484 5. The Makefile now contains an "install" target, with editable destinations at
5485 the top of the file. The pcretest program is not installed.
5487 6. pgrep -V now gives the PCRE version number and date.
5489 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5490 causing the entire string to be ignored, instead of just the last character.
5492 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5493 non-matching string, it can take a very, very long time, even for strings of
5494 quite modest length, because of the nested recursion. PCRE now does better in
5495 some of these cases. It does this by remembering the last required literal
5496 character in the pattern, and pre-searching the subject to ensure it is present
5497 before running the real match. In other words, it applies a heuristic to detect
5498 some types of certain failure quickly, and in the above example, if presented
5499 with a string that has no trailing " it gives "no match" very quickly.
5501 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5502 other alternatives are tried instead.
5505 Version 2.06 09-Jun-99
5506 ----------------------
5508 1. Change pcretest's output for amount of store used to show just the code
5509 space, because the remainder (the data block) varies in size between 32-bit and
5510 64-bit systems.
5512 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5513 start matching at. This allows lookbehinds to work when searching for multiple
5514 occurrences in a string.
5516 3. Added additional options to pcretest for testing multiple occurrences:
5518 /+ outputs the rest of the string that follows a match
5519 /g loops for multiple occurrences, using the new startoffset argument
5520 /G loops for multiple occurrences by passing an incremented pointer
5522 4. PCRE wasn't doing the "first character" optimization for patterns starting
5523 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5524 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5525 the letter 'x'. On long subject strings, this gives a significant speed-up.
5528 Version 2.05 21-Apr-99
5529 ----------------------
5531 1. Changed the type of magic_number from int to long int so that it works
5532 properly on 16-bit systems.
5534 2. Fixed a bug which caused patterns starting with .* not to work correctly
5535 when the subject string contained newline characters. PCRE was assuming
5536 anchoring for such patterns in all cases, which is not correct because .* will
5537 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5538 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5539 must be retried after every newline in the subject.
5542 Version 2.04 18-Feb-99
5543 ----------------------
5545 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5546 computation of the store needed to hold the pattern was incorrect (too large).
5547 If such patterns were nested a few deep, this could multiply and become a real
5548 problem.
5550 2. Added /M option to pcretest to show the memory requirement of a specific
5551 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5553 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5554 compiled in such a way that the backtracking after subsequent failure was
5555 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5556 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5559 Version 2.03 02-Feb-99
5560 ----------------------
5562 1. Fixed typo and small mistake in man page.
5564 2. Added 4th condition (GPL supersedes if conflict) and created separate
5565 LICENCE file containing the conditions.
5567 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5568 Perl, that is the internal \ allows the delimiter to be included in the
5569 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5570 the final delimiter, add \ to the end of the pattern (to test the error).
5572 4. Added the convenience functions for extracting substrings after a successful
5573 match. Updated pcretest to make it able to test these functions.
5576 Version 2.02 14-Jan-99
5577 ----------------------
5579 1. Initialized the working variables associated with each extraction so that
5580 their saving and restoring doesn't refer to uninitialized store.
5582 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5583 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5584 fix the problem.
5586 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5587 calls, and wasn't printing the correct value for compiling calls. Increased the
5588 default value of LOOPREPEAT, and the number of significant figures in the
5589 times.
5591 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5593 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5594 a building problem on Windows NT with a FAT file system.
5597 Version 2.01 21-Oct-98
5598 ----------------------
5600 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5601 to character tables built by pcre_maketables() in the current locale. If NULL
5602 is passed, the default tables are used.
5605 Version 2.00 24-Sep-98
5606 ----------------------
5608 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5609 it any more.
5611 2. Allow quantification of (?>) groups, and make it work correctly.
5613 3. The first character computation wasn't working for (?>) groups.
5615 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5616 end of the subject) and add 5.005's \z, which really does match only at the
5617 very end of the subject.
5619 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5621 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5622 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5623 localized options. All options to pcre_study() were also removed.
5625 7. Add other new features from 5.005:
5627 $(?<= positive lookbehind
5628 $(?<! negative lookbehind
5629 (?imsx-imsx) added the unsetting capability
5630 such a setting is global if at outer level; local otherwise
5631 (?imsx-imsx:) non-capturing groups with option setting
5632 (?(cond)re|re) conditional pattern matching
5634 A backreference to itself in a repeated group matches the previous
5635 captured string.
5637 8. General tidying up of studying (both automatic and via "study")
5638 consequential on the addition of new assertions.
5640 9. As in 5.005, unlimited repeated groups that could match an empty substring
5641 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5642 runtime if any iteration does actually match an empty substring.
5644 10. Include the RunTest script in the distribution.
5646 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5647 discrepancies, some of which were old and were also with respect to 5.004. They
5648 have now been fixed.
5651 Version 1.09 28-Apr-98
5652 ----------------------
5654 1. A negated single character class followed by a quantifier with a minimum
5655 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5656 program crashes, or just wrong answers. This did not apply to negated classes
5657 containing more than one character, or to minima other than one.
5660 Version 1.08 27-Mar-98
5661 ----------------------
5663 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5665 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5666 latter must appear before anything that relies on it in the pattern.
5669 Version 1.07 16-Feb-98
5670 ----------------------
5672 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5673 repeat of a potentially empty string).
5676 Version 1.06 23-Jan-98
5677 ----------------------
5679 1. Added Markus Oberhumer's little patches for C++.
5681 2. Literal strings longer than 255 characters were broken.
5684 Version 1.05 23-Dec-97
5685 ----------------------
5687 1. Negated character classes containing more than one character were failing if
5688 PCRE_CASELESS was set at run time.
5691 Version 1.04 19-Dec-97
5692 ----------------------
5694 1. Corrected the man page, where some "const" qualifiers had been omitted.
5696 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5697 input syntax.
5699 3. Fixed memory leak which occurred when a regex with back references was
5700 matched with an offsets vector that wasn't big enough. The temporary memory
5701 that is used in this case wasn't being freed if the match failed.
5703 4. Tidied pcretest to ensure it frees memory that it gets.
5705 5. Temporary memory was being obtained in the case where the passed offsets
5706 vector was exactly big enough.
5708 6. Corrected definition of offsetof() from change 5 below.
5710 7. I had screwed up change 6 below and broken the rules for the use of
5711 setjmp(). Now fixed.
5714 Version 1.03 18-Dec-97
5715 ----------------------
5717 1. A erroneous regex with a missing opening parenthesis was correctly
5718 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5719 on some systems.
5721 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5722 it was reported that one broken compiler failed on the former because "code" is
5723 also an independent variable.
5725 3. The erroneous regex a[]b caused an array overrun reference.
5727 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5728 fail on data ending with that character. (It was going on too far, and checking
5729 the next character, typically a binary zero.) This was specific to the
5730 optimized code for single-character negative classes.
5732 5. Added a contributed patch from the TIN world which does the following:
5734 + Add an undef for memmove, in case the the system defines a macro for it.
5736 + Add a definition of offsetof(), in case there isn't one. (I don't know
5737 the reason behind this - offsetof() is part of the ANSI standard - but
5738 it does no harm).
5740 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5741 most of the places where whitespace preceded '#'. I have given up and
5742 allowed the remaining 2 cases to be at the margin.
5744 + Rename some variables in pcre to eliminate shadowing. This seems very
5745 pedantic, but does no harm, of course.
5747 6. Moved the call to setjmp() into its own function, to get rid of warnings
5748 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5750 7. Constructs such as \d{8,} were compiling into the equivalent of
5751 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5752 outcome, but in this particular case used more store than had been allocated,
5753 which caused the bug to be discovered because it threw up an internal error.
5755 8. The debugging code in both pcre and pcretest for outputting the compiled
5756 form of a regex was going wrong in the case of back references followed by
5757 curly-bracketed repeats.
5760 Version 1.02 12-Dec-97
5761 ----------------------
5763 1. Typos in pcre.3 and comments in the source fixed.
5765 2. Applied a contributed patch to get rid of places where it used to remove
5766 'const' from variables, and fixed some signed/unsigned and uninitialized
5767 variable warnings.
5769 3. Added the "runtest" target to Makefile.
5771 4. Set default compiler flag to -O2 rather than just -O.
5774 Version 1.01 19-Nov-97
5775 ----------------------
5777 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5778 like /([ab]*)*/, that is, for classes with more than one character in them.
5780 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5781 as /((?>a*))*/ (a PCRE_EXTRA facility).
5784 Version 1.00 18-Nov-97
5785 ----------------------
5787 1. Added compile-time macros to support systems such as SunOS4 which don't have
5788 memmove() or strerror() but have other things that can be used instead.
5790 2. Arranged that "make clean" removes the executables.
5793 Version 0.99 27-Oct-97
5794 ----------------------
5796 1. Fixed bug in code for optimizing classes with only one character. It was
5797 initializing a 32-byte map regardless, which could cause it to run off the end
5798 of the memory it had got.
5800 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5803 Version 0.98 22-Oct-97
5804 ----------------------
5806 1. Fixed bug in code for handling temporary memory usage when there are more
5807 back references than supplied space in the ovector. This could cause segfaults.
5810 Version 0.97 21-Oct-97
5811 ----------------------
5813 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5815 2. Optimized negated single characters not to use a bit map.
5817 3. Brought error texts together as macro definitions; clarified some of them;
5818 fixed one that was wrong - it said "range out of order" when it meant "invalid
5819 escape sequence".
5821 4. Changed some char * arguments to const char *.
5823 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5825 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5826 pcretest.
5829 Version 0.96 16-Oct-97
5830 ----------------------
5832 1. Added a simple "pgrep" utility to the distribution.
5834 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5835 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5836 where "ddd" means "one or more decimal digits".
5838 3. Fixed serious bug. If a pattern had a back reference, but the call to
5839 pcre_exec() didn't supply a large enough ovector to record the related
5840 identifying subpattern, the match always failed. PCRE now remembers the number
5841 of the largest back reference, and gets some temporary memory in which to save
5842 the offsets during matching if necessary, in order to ensure that
5843 backreferences always work.
5845 4. Increased the compatibility with Perl in a number of ways:
5847 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5848 to request this handling. The option can be set at compile or exec time.
5850 (b) $ matches before a terminating newline by default; an option
5851 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5852 mode). The option can be set at compile or exec time.
5854 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5855 the same as Perl's. If the decimal number it represents is less than 10
5856 or there aren't that many previous left capturing parentheses, an octal
5857 escape is read. Inside a character class, it's always an octal escape,
5858 even if it is a single digit.
5860 (d) An escaped but undefined alphabetic character is taken as a literal,
5861 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5862 escapes.
5864 (e) {0} is now permitted. (The previous item is removed from the compiled
5865 pattern).
5867 5. Changed all the names of code files so that the basic parts are no longer
5868 than 10 characters, and abolished the teeny "globals.c" file.
5870 6. Changed the handling of character classes; they are now done with a 32-byte
5871 bit map always.
5873 7. Added the -d and /D options to pcretest to make it possible to look at the
5874 internals of compilation without having to recompile pcre.
5877 Version 0.95 23-Sep-97
5878 ----------------------
5880 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5881 \x20 at the start of a run of normal characters. These were being treated as
5882 real characters, instead of the source characters being re-checked.
5885 Version 0.94 18-Sep-97
5886 ----------------------
5888 1. The functions are now thread-safe, with the caveat that the global variables
5889 containing pointers to malloc() and free() or alternative functions are the
5890 same for all threads.
5892 2. Get pcre_study() to generate a bitmap of initial characters for non-
5893 anchored patterns when this is possible, and use it if passed to pcre_exec().
5896 Version 0.93 15-Sep-97
5897 ----------------------
5899 1. /(b)|(:+)/ was computing an incorrect first character.
5901 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5902 but not actually doing anything yet.
5904 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5905 as Perl does (e.g. [-az] or [az-]).
5907 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5908 all possible positions.
5910 5. Split up into different modules to avoid including unneeded functions in a
5911 compiled binary. However, compile and exec are still in one module. The "study"
5912 function is split off.
5914 6. The character tables are now in a separate module whose source is generated
5915 by an auxiliary program - but can then be edited by hand if required. There are
5916 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5917 toupper() in the code.
5919 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5920 make them global. Abolish the function for setting them, as the caller can now
5921 set them directly.
5924 Version 0.92 11-Sep-97
5925 ----------------------
5927 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5928 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5930 2. Caseless matching was not working in character classes if the characters in
5931 the pattern were in upper case.
5933 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5935 4. Make PCRE_ANCHORED public and accept as a compile option.
5937 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5938 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5939 pass them.
5941 6. Give an error if bad option bits passed at compile or run time.
5943 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5944 pcretest to cause it to pass that flag.
5946 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5947 options, and the first character, if set.
5949 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5952 Version 0.91 10-Sep-97
5953 ----------------------
5955 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5956 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5958 2. PCRE was looping on encountering an indefinitely repeated back reference to
5959 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5960 Perl does - treats the match as successful.
5962 ****


Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5