ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1612 - (show annotations)
Fri Nov 27 17:13:13 2015 UTC (3 years, 10 months ago) by ph10
File size: 285834 byte(s)
Fix negated POSIX class within negated overall class UCP bug.
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. Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
15 working correctly in UCP mode.
18 Version 8.38 23-November-2015
19 -----------------------------
21 1. If a group that contained a recursive back reference also contained a
22 forward reference subroutine call followed by a non-forward-reference
23 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
24 compile correct code, leading to undefined behaviour or an internally
25 detected error. This bug was discovered by the LLVM fuzzer.
27 2. Quantification of certain items (e.g. atomic back references) could cause
28 incorrect code to be compiled when recursive forward references were
29 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
30 This bug was discovered by the LLVM fuzzer.
32 3. A repeated conditional group whose condition was a reference by name caused
33 a buffer overflow if there was more than one group with the given name.
34 This bug was discovered by the LLVM fuzzer.
36 4. A recursive back reference by name within a group that had the same name as
37 another group caused a buffer overflow. For example:
38 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
40 5. A forward reference by name to a group whose number is the same as the
41 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
42 a buffer overflow at compile time. This bug was discovered by the LLVM
43 fuzzer.
45 6. A lookbehind assertion within a set of mutually recursive subpatterns could
46 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
48 7. Another buffer overflow bug involved duplicate named groups with a
49 reference between their definition, with a group that reset capture
50 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
51 fixed by always allowing for more memory, even if not needed. (A proper fix
52 is implemented in PCRE2, but it involves more refactoring.)
54 8. There was no check for integer overflow in subroutine calls such as (?123).
56 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
57 being treated as a literal 'l' instead of causing an error.
59 10. There was a buffer overflow if pcre_exec() was called with an ovector of
60 size 1. This bug was found by american fuzzy lop.
62 11. If a non-capturing group containing a conditional group that could match
63 an empty string was repeated, it was not identified as matching an empty
64 string itself. For example: /^(?:(?(1)x|)+)+$()/.
66 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
67 \a and \e in test subject lines.
69 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
70 instead of the EBCDIC value.
72 14. The handling of \c in an EBCDIC environment has been revised so that it is
73 now compatible with the specification in Perl's perlebcdic page.
75 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
76 ASCII/Unicode. This has now been added to the list of characters that are
77 recognized as white space in EBCDIC.
79 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
80 error (correctly) when used outside a class, but did not give an error
81 within a class.
83 17. \h within a class was incorrectly compiled in EBCDIC environments.
85 18. A pattern with an unmatched closing parenthesis that contained a backward
86 assertion which itself contained a forward reference caused buffer
87 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
89 19. JIT should return with error when the compiled pattern requires more stack
90 space than the maximum.
92 20. A possessively repeated conditional group that could match an empty string,
93 for example, /(?(R))*+/, was incorrectly compiled.
95 21. Fix infinite recursion in the JIT compiler when certain patterns such as
96 /(?:|a|){100}x/ are analysed.
98 22. Some patterns with character classes involving [: and \\ were incorrectly
99 compiled and could cause reading from uninitialized memory or an incorrect
100 error diagnosis.
102 23. Pathological patterns containing many nested occurrences of [: caused
103 pcre_compile() to run for a very long time.
105 24. A conditional group with only one branch has an implicit empty alternative
106 branch and must therefore be treated as potentially matching an empty
107 string.
109 25. If (?R was followed by - or + incorrect behaviour happened instead of a
110 diagnostic.
112 26. Arrange to give up on finding the minimum matching length for overly
113 complex patterns.
115 27. Similar to (4) above: in a pattern with duplicated named groups and an
116 occurrence of (?| it is possible for an apparently non-recursive back
117 reference to become recursive if a later named group with the relevant
118 number is encountered. This could lead to a buffer overflow. Wen Guanxing
119 from Venustech ADLAB discovered this bug.
121 28. If pcregrep was given the -q option with -c or -l, or when handling a
122 binary file, it incorrectly wrote output to stdout.
124 29. The JIT compiler did not restore the control verb head in case of *THEN
125 control verbs. This issue was found by Karl Skomski with a custom LLVM
126 fuzzer.
128 30. Error messages for syntax errors following \g and \k were giving inaccurate
129 offsets in the pattern.
131 31. Added a check for integer overflow in conditions (?(<digits>) and
132 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
133 fuzzer.
135 32. Handling recursive references such as (?2) when the reference is to a group
136 later in the pattern uses code that is very hacked about and error-prone.
137 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
138 give an internal error if it is obvious that compiling has gone wrong.
140 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
141 This issue was found by Karl Skomski with a custom LLVM fuzzer.
143 34. The JIT compiler should restore the control chain for empty possessive
144 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
146 35. Match limit check added to JIT recursion. This issue was found by Karl
147 Skomski with a custom LLVM fuzzer.
149 36. Yet another case similar to 27 above has been circumvented by an
150 unconditional allocation of extra memory. This issue is fixed "properly" in
151 PCRE2 by refactoring the way references are handled. Wen Guanxing
152 from Venustech ADLAB discovered this bug.
154 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
155 with a custom LLVM fuzzer.
157 38. Fixed a corner case of range optimization in JIT.
159 39. An incorrect error "overran compiling workspace" was given if there were
160 exactly enough group forward references such that the last one extended
161 into the workspace safety margin. The next one would have expanded the
162 workspace. The test for overflow was not including the safety margin.
164 40. A match limit issue is fixed in JIT which was found by Karl Skomski
165 with a custom LLVM fuzzer.
167 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
168 work under Windows. (Why has it taken so long for anyone to notice?)
170 42. In a character class such as [\W\p{Any}] where both a negative-type escape
171 ("not a word character") and a property escape were present, the property
172 escape was being ignored.
174 43. Fix crash caused by very long (*MARK) or (*THEN) names.
176 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
177 by a single ASCII character in a class item, was incorrectly compiled in
178 UCP mode. The POSIX class got lost, but only if the single character
179 followed it.
181 45. [:punct:] in UCP mode was matching some characters in the range 128-255
182 that should not have been matched.
184 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
185 class, all characters with code points greater than 255 are in the class.
186 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
187 such as \w are turned into Unicode properties), wide characters were not
188 correctly handled, and could fail to match.
191 Version 8.37 28-April-2015
192 --------------------------
194 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
195 for those parentheses to be closed with whatever has been captured so far.
196 However, it was failing to mark any other groups between the hightest
197 capture so far and the currrent group as "unset". Thus, the ovector for
198 those groups contained whatever was previously there. An example is the
199 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
201 2. If an assertion condition was quantified with a minimum of zero (an odd
202 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
204 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
205 unrecognized modifier, a crash could occur.
207 4. An attempt to do global matching in pcretest with a zero-length ovector
208 caused a crash.
210 5. Fixed a memory leak during matching that could occur for a subpattern
211 subroutine call (recursive or otherwise) if the number of captured groups
212 that had to be saved was greater than ten.
214 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
215 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
216 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
218 7. A UTF pattern containing a "not" match of a non-ASCII character and a
219 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
221 8. When a pattern is compiled, it remembers the highest back reference so that
222 when matching, if the ovector is too small, extra memory can be obtained to
223 use instead. A conditional subpattern whose condition is a check on a
224 capture having happened, such as, for example in the pattern
225 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
226 setting the highest backreference number. This mattered only if pcre_exec()
227 was called with an ovector that was too small to hold the capture, and there
228 was no other kind of back reference (a situation which is probably quite
229 rare). The effect of the bug was that the condition was always treated as
230 FALSE when the capture could not be consulted, leading to a incorrect
231 behaviour by pcre_exec(). This bug has been fixed.
233 9. A reference to a duplicated named group (either a back reference or a test
234 for being set in a conditional) that occurred in a part of the pattern where
235 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
236 to be incorrectly calculated, leading to overwriting.
238 10. A mutually recursive set of back references such as (\2)(\1) caused a
239 segfault at study time (while trying to find the minimum matching length).
240 The infinite loop is now broken (with the minimum length unset, that is,
241 zero).
243 11. If an assertion that was used as a condition was quantified with a minimum
244 of zero, matching went wrong. In particular, if the whole group had
245 unlimited repetition and could match an empty string, a segfault was
246 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
247 assertions to be quantified, but not if they are being used as conditions,
248 so the above pattern is faulted by Perl. PCRE has now been changed so that
249 it also rejects such patterns.
251 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
252 failed to allow the zero-repeat case if pcre2_exec() was called with an
253 ovector too small to capture the group.
255 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
256 Red Hat Product Security:
258 (a) A crash if /K and /F were both set with the option to save the compiled
259 pattern.
261 (b) Another crash if the option to print captured substrings in a callout
262 was combined with setting a null ovector, for example \O\C+ as a subject
263 string.
265 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
266 forward reference repeated a large (but limited) number of times within a
267 repeated outer group that has a zero minimum quantifier, caused incorrect
268 code to be compiled, leading to the error "internal error:
269 previously-checked referenced subpattern not found" when an incorrect
270 memory address was read. This bug was reported as "heap overflow",
271 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
272 CVE-2015-2325.
274 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
275 call within a group that also contained a recursive back reference caused
276 incorrect code to be compiled. This bug was reported as "heap overflow",
277 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
278 number CVE-2015-2326.
280 24. Computing the size of the JIT read-only data in advance has been a source
281 of various issues, and new ones are still appear unfortunately. To fix
282 existing and future issues, size computation is eliminated from the code,
283 and replaced by on-demand memory allocation.
285 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
286 adjacent to the end of the range, and the range contained characters with
287 more than one other case, caused incorrect behaviour when compiled in UTF
288 mode. In that example, the range a-j was left out of the class.
290 26. Fix JIT compilation of conditional blocks, which assertion
291 is converted to (*FAIL). E.g: /(?(?!))/.
293 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
294 discovered by the LLVM fuzzer.
296 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
297 when this assertion was used as a condition, for example (?(?!)a|b). In
298 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
299 error about an unsupported item.
301 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
302 possessification code could take exponential time to complete. A recursion
303 depth limit of 1000 has been imposed to limit the resources used by this
304 optimization.
306 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
307 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
308 because \S ensures they are all in the class. The code for doing this was
309 interacting badly with the code for computing the amount of space needed to
310 compile the pattern, leading to a buffer overflow. This bug was discovered
311 by the LLVM fuzzer.
313 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
314 other kinds of group caused stack overflow at compile time. This bug was
315 discovered by the LLVM fuzzer.
317 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
318 between a subroutine call and its quantifier was incorrectly compiled,
319 leading to buffer overflow or other errors. This bug was discovered by the
320 LLVM fuzzer.
322 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
323 assertion after (?(. The code was failing to check the character after
324 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
325 was discovered by the LLVM fuzzer.
327 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
328 a fixed maximum following a group that contains a subroutine reference was
329 incorrectly compiled and could trigger buffer overflow. This bug was
330 discovered by the LLVM fuzzer.
332 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
333 caused a stack overflow instead of the diagnosis of a non-fixed length
334 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
336 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
337 (e.g. /(?<=\Ka)/) could make pcregrep loop.
339 37. There was a similar problem to 36 in pcretest for global matches.
341 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
342 and a subsequent item in the pattern caused a non-match, backtracking over
343 the repeated \X did not stop, but carried on past the start of the subject,
344 causing reference to random memory and/or a segfault. There were also some
345 other cases where backtracking after \C could crash. This set of bugs was
346 discovered by the LLVM fuzzer.
348 39. The function for finding the minimum length of a matching string could take
349 a very long time if mutual recursion was present many times in a pattern,
350 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
351 method has been implemented. This infelicity was discovered by the LLVM
352 fuzzer.
354 40. Static linking against the PCRE library using the pkg-config module was
355 failing on missing pthread symbols.
358 Version 8.36 26-September-2014
359 ------------------------------
361 1. Got rid of some compiler warnings in the C++ modules that were shown up by
362 -Wmissing-field-initializers and -Wunused-parameter.
364 2. The tests for quantifiers being too big (greater than 65535) were being
365 applied after reading the number, and stupidly assuming that integer
366 overflow would give a negative number. The tests are now applied as the
367 numbers are read.
369 3. Tidy code in pcre_exec.c where two branches that used to be different are
370 now the same.
372 4. The JIT compiler did not generate match limit checks for certain
373 bracketed expressions with quantifiers. This may lead to exponential
374 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
375 issue should be resolved now.
377 5. Fixed an issue, which occures when nested alternatives are optimized
378 with table jumps.
380 6. Inserted two casts and changed some ints to size_t in the light of some
381 reported 64-bit compiler warnings (Bugzilla 1477).
383 7. Fixed a bug concerned with zero-minimum possessive groups that could match
384 an empty string, which sometimes were behaving incorrectly in the
385 interpreter (though correctly in the JIT matcher). This pcretest input is
386 an example:
388 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
391 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
392 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
393 for an empty string was breaking the inner loop and carrying on at a lower
394 level, when possessive repeated groups should always return to a higher
395 level as they have no backtrack points in them. The empty string test now
396 occurs at the outer level.
398 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
399 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
401 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
402 doesn't).
404 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
405 to hold a pattern to be incorrectly computed (too small) when there were
406 named back references to duplicated names. This could cause "internal
407 error: code overflow" or "double free or corruption" or other memory
408 handling errors.
410 11. When named subpatterns had the same prefixes, back references could be
411 confused. For example, in this pattern:
413 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
415 the reference to 'Name' was incorrectly treated as a reference to a
416 duplicate name.
418 12. A pattern such as /^s?c/mi8 where the optional character has more than
419 one "other case" was incorrectly compiled such that it would only try to
420 match starting at "c".
422 13. When a pattern starting with \s was studied, VT was not included in the
423 list of possible starting characters; this should have been part of the
424 8.34/18 patch.
426 14. If a character class started [\Qx]... where x is any character, the class
427 was incorrectly terminated at the ].
429 15. If a pattern that started with a caseless match for a character with more
430 than one "other case" was studied, PCRE did not set up the starting code
431 unit bit map for the list of possible characters. Now it does. This is an
432 optimization improvement, not a bug fix.
434 16. The Unicode data tables have been updated to Unicode 7.0.0.
436 17. Fixed a number of memory leaks in pcregrep.
438 18. Avoid a compiler warning (from some compilers) for a function call with
439 a cast that removes "const" from an lvalue by using an intermediate
440 variable (to which the compiler does not object).
442 19. Incorrect code was compiled if a group that contained an internal recursive
443 back reference was optional (had quantifier with a minimum of zero). This
444 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
445 caused segmentation faults because of stack overflows at compile time.
447 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
448 group that is quantified with an indefinite repeat, caused a compile-time
449 loop which used up all the system stack and provoked a segmentation fault.
450 This was not the same bug as 19 above.
452 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
453 Patch by Mike Frysinger.
456 Version 8.35 04-April-2014
457 --------------------------
459 1. A new flag is set, when property checks are present in an XCLASS.
460 When this flag is not set, PCRE can perform certain optimizations
461 such as studying these XCLASS-es.
463 2. The auto-possessification of character sets were improved: a normal
464 and an extended character set can be compared now. Furthermore
465 the JIT compiler optimizes more character set checks.
467 3. Got rid of some compiler warnings for potentially uninitialized variables
468 that show up only when compiled with -O2.
470 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
471 of a match later then the end of the match. The pcretest program was not
472 handling the case sensibly - it was outputting from the start to the next
473 binary zero. It now reports this situation in a message, and outputs the
474 text from the end to the start.
476 5. Fast forward search is improved in JIT. Instead of the first three
477 characters, any three characters with fixed position can be searched.
478 Search order: first, last, middle.
480 6. Improve character range checks in JIT. Characters are read by an inprecise
481 function now, which returns with an unknown value if the character code is
482 above a certain threshold (e.g: 256). The only limitation is that the value
483 must be bigger than the threshold as well. This function is useful when
484 the characters above the threshold are handled in the same way.
486 7. The macros whose names start with RAWUCHAR are placeholders for a future
487 mode in which only the bottom 21 bits of 32-bit data items are used. To
488 make this more memorable for those maintaining the code, the names have
489 been changed to start with UCHAR21, and an extensive comment has been added
490 to their definition.
492 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
493 to the export list in Makefile.am (they were accidentally omitted from the
494 8.34 tarball).
496 9. The informational output from pcretest used the phrase "starting byte set"
497 which is inappropriate for the 16-bit and 32-bit libraries. As the output
498 for "first char" and "need char" really means "non-UTF-char", I've changed
499 "byte" to "char", and slightly reworded the output. The documentation about
500 these values has also been (I hope) clarified.
502 10. Another JIT related optimization: use table jumps for selecting the correct
503 backtracking path, when more than four alternatives are present inside a
504 bracket.
506 11. Empty match is not possible, when the minimum length is greater than zero,
507 and there is no \K in the pattern. JIT should avoid empty match checks in
508 such cases.
510 12. In a caseless character class with UCP support, when a character with more
511 than one alternative case was not the first character of a range, not all
512 the alternative cases were added to the class. For example, s and \x{17f}
513 are both alternative cases for S: the class [RST] was handled correctly,
514 but [R-T] was not.
516 13. The configure.ac file always checked for pthread support when JIT was
517 enabled. This is not used in Windows, so I have put this test inside a
518 check for the presence of windows.h (which was already tested for).
520 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
521 The algorithm provides a way to skip certain starting offsets, and usually
522 faster than linear prefix searches.
524 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
525 as for 'fr_FR' and 'french'. For some reason, however, it then used the
526 Windows-specific input and output files, which have 'french' screwed in.
527 So this could never have worked. One of the problems with locales is that
528 they aren't always the same. I have now updated RunTest so that it checks
529 the output of the locale test (test 3) against three different output
530 files, and it allows the test to pass if any one of them matches. With luck
531 this should make the test pass on some versions of Solaris where it was
532 failing. Because of the uncertainty, the script did not used to stop if
533 test 3 failed; it now does. If further versions of a French locale ever
534 come to light, they can now easily be added.
536 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
537 there was a message during ./configure, but it did not stop. This now
538 provokes an error. The invalid example in README has been corrected.
539 If a value less than the minimum is given, the minimum value has always
540 been used, but now a warning is given.
542 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
543 was a bug in the test system, which is now fixed. Also, the list of various
544 configurations that are tested for each release did not have one with both
545 16/32 bits and --enable-bar-anycrlf. It now does.
547 18. pcretest was missing "-C bsr" for displaying the \R default setting.
549 19. Little endian PowerPC systems are supported now by the JIT compiler.
551 20. The fast forward newline mechanism could enter to an infinite loop on
552 certain invalid UTF-8 input. Although we don't support these cases
553 this issue can be fixed by a performance optimization.
555 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
556 not take account if existing stack usage. There is now a new global
557 variable called pcre_stack_guard that can be set to point to an external
558 function to check stack availability. It is called at the start of
559 processing every parenthesized group.
561 22. A typo in the code meant that in ungreedy mode the max/min qualifier
562 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
563 not match "ab".
565 23. When UTF was disabled, the JIT program reported some incorrect compile
566 errors. These messages are silenced now.
568 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
569 compiler.
571 25. Change all the temporary files used in RunGrepTest to be different to those
572 used by RunTest so that the tests can be run simultaneously, for example by
573 "make -j check".
576 Version 8.34 15-December-2013
577 -----------------------------
579 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
580 executable memory. Patch inspired by Carsten Klein.
582 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
583 this macro is never tested and has no effect, because the work to support
584 coverage involves only compiling and linking options and special targets in
585 the Makefile. The comment in config.h implied that defining the macro would
586 enable coverage support, which is totally false. There was also support for
587 setting this macro in the CMake files (my fault, I just copied it from
588 configure). SUPPORT_GCOV has now been removed.
590 3. Make a small performance improvement in strlen16() and strlen32() in
591 pcretest.
593 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
594 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
595 these cases). There was also one in pcretest.c.
597 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
599 6. In UTF mode, the code for checking whether a group could match an empty
600 string (which is used for indefinitely repeated groups to allow for
601 breaking an infinite loop) was broken when the group contained a repeated
602 negated single-character class with a character that occupied more than one
603 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
604 UTF-8 mode). The effect was undefined: the group might or might not be
605 deemed as matching an empty string, or the program might have crashed.
607 7. The code for checking whether a group could match an empty string was not
608 recognizing that \h, \H, \v, \V, and \R must match a character.
610 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
611 an empty string. If it can, pcretest shows this in its information output.
613 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
614 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
615 matched by pcre_exec() without using JIT:
617 (a) If the rest of the pattern did not match after a maximal run of
618 grapheme clusters, the code for backing up to try with fewer of them
619 did not always back up over a full grapheme when characters that do not
620 have the modifier quality were involved, e.g. Hangul syllables.
622 (b) If the match point in a subject started with modifier character, and
623 there was no match, the code could incorrectly back up beyond the match
624 point, and potentially beyond the first character in the subject,
625 leading to a segfault or an incorrect match result.
627 10. A conditional group with an assertion condition could lead to PCRE
628 recording an incorrect first data item for a match if no other first data
629 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
630 first data item, and therefore matched "ca" after "c" instead of at the
631 start.
633 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
634 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
636 12. The source of pcregrep now includes z/OS-specific code so that it can be
637 compiled for z/OS as part of the special z/OS distribution.
639 13. Added the -T and -TM options to pcretest.
641 14. The code in pcre_compile.c for creating the table of named capturing groups
642 has been refactored. Instead of creating the table dynamically during the
643 actual compiling pass, the information is remembered during the pre-compile
644 pass (on the stack unless there are more than 20 named groups, in which
645 case malloc() is used) and the whole table is created before the actual
646 compile happens. This has simplified the code (it is now nearly 150 lines
647 shorter) and prepared the way for better handling of references to groups
648 with duplicate names.
650 15. A back reference to a named subpattern when there is more than one of the
651 same name now checks them in the order in which they appear in the pattern.
652 The first one that is set is used for the reference. Previously only the
653 first one was inspected. This change makes PCRE more compatible with Perl.
655 16. Unicode character properties were updated from Unicode 6.3.0.
657 17. The compile-time code for auto-possessification has been refactored, based
658 on a patch by Zoltan Herczeg. It now happens after instead of during
659 compilation. The code is cleaner, and more cases are handled. The option
660 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
661 options in pcretest are provided to set it. It can also be set by
662 (*NO_AUTO_POSSESS) at the start of a pattern.
664 18. The character VT has been added to the default ("C" locale) set of
665 characters that match \s and are generally treated as white space,
666 following this same change in Perl 5.18. There is now no difference between
667 "Perl space" and "POSIX space". Whether VT is treated as white space in
668 other locales depends on the locale.
670 19. The code for checking named groups as conditions, either for being set or
671 for being recursed, has been refactored (this is related to 14 and 15
672 above). Processing unduplicated named groups should now be as fast at
673 numerical groups, and processing duplicated groups should be faster than
674 before.
676 20. Two patches to the CMake build system, by Alexander Barkov:
678 (1) Replace the "source" command by "." in CMakeLists.txt because
679 "source" is a bash-ism.
681 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
682 without these the CMake build does not work on Solaris.
684 21. Perl has changed its handling of \8 and \9. If there is no previously
685 encountered capturing group of those numbers, they are treated as the
686 literal characters 8 and 9 instead of a binary zero followed by the
687 literals. PCRE now does the same.
689 22. Following Perl, added \o{} to specify codepoints in octal, making it
690 possible to specify values greater than 0777 and also making them
691 unambiguous.
693 23. Perl now gives an error for missing closing braces after \x{... instead of
694 treating the string as literal. PCRE now does the same.
696 24. RunTest used to grumble if an inappropriate test was selected explicitly,
697 but just skip it when running all tests. This make it awkward to run ranges
698 of tests when one of them was inappropriate. Now it just skips any
699 inappropriate tests, as it always did when running all tests.
701 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
702 character types such as \d or \w, too many callouts were inserted, and the
703 data that they returned was rubbish.
705 26. In UCP mode, \s was not matching two of the characters that Perl matches,
706 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
707 were matched by \h. The code has now been refactored so that the lists of
708 the horizontal and vertical whitespace characters used for \h and \v (which
709 are defined only in one place) are now also used for \s.
711 27. Add JIT support for the 64 bit TileGX architecture.
712 Patch by Jiong Wang (Tilera Corporation).
714 28. Possessive quantifiers for classes (both explicit and automatically
715 generated) now use special opcodes instead of wrapping in ONCE brackets.
717 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
718 (because it's meaningless), this was not happening when PCRE_CASELESS was
719 set. Not wrong, but inefficient.
721 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
722 properties for \w, \d, etc) is present in a test regex. Otherwise if the
723 test contains no characters greater than 255, Perl doesn't realise it
724 should be using Unicode semantics.
726 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
727 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
728 does in Unicode mode.
730 32. Added the "forbid" facility to pcretest so that putting tests into the
731 wrong test files can sometimes be quickly detected.
733 33. There is now a limit (default 250) on the depth of nesting of parentheses.
734 This limit is imposed to control the amount of system stack used at compile
735 time. It can be changed at build time by --with-parens-nest-limit=xxx or
736 the equivalent in CMake.
738 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
739 errors. Perl warns for these when in warning mode, but PCRE has no facility
740 for giving warnings.
742 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
743 However, this was not working for (?!) because it is optimized to (*FAIL),
744 for which PCRE does not allow quantifiers. The optimization is now disabled
745 when a quantifier follows (?!). I can't see any use for this, but it makes
746 things uniform.
748 36. Perl no longer allows group names to start with digits, so I have made this
749 change also in PCRE. It simplifies the code a bit.
751 37. In extended mode, Perl ignores spaces before a + that indicates a
752 possessive quantifier. PCRE allowed a space before the quantifier, but not
753 before the possessive +. It now does.
755 38. The use of \K (reset reported match start) within a repeated possessive
756 group such as (a\Kb)*+ was not working.
758 40. Document that the same character tables must be used at compile time and
759 run time, and that the facility to pass tables to pcre_exec() and
760 pcre_dfa_exec() is for use only with saved/restored patterns.
762 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
763 features for Builds with MSVC:
765 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
766 builds). The .rc files can be used to set FileDescription and many other
767 attributes.
769 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
770 This allows higher-level build scripts which want .pdb files to avoid
771 hard-coding the exact files needed."
773 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
774 mean "start of word" and "end of word", respectively, as a transition aid.
776 43. A minimizing repeat of a class containing codepoints greater than 255 in
777 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
778 compiled to use the heap for recursion.
780 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
781 is configured.
784 Version 8.33 28-May-2013
785 ------------------------
787 1. Added 'U' to some constants that are compared to unsigned integers, to
788 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
789 variables that are added to signed variables, to ensure the result is
790 signed and can be negated.
792 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
793 CMake config files.
795 3. Revise the creation of config.h.generic so that all boolean macros are
796 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
797 overriding via -D on the command line possible.
799 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
800 to unsigned int is reported to make a quite noticeable speed difference in
801 a specific Windows environment. Testing on Linux did also appear to show
802 some benefit (and it is clearly not harmful). Also fixed the definition of
803 Xop which should be unsigned.
805 5. Related to (4), changing the definition of the intermediate variable cc
806 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
807 improvements.
809 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
810 unnecessary spaces.
812 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
813 and later.
815 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
817 9. Optimizing fast_forward_start_bits in JIT.
819 10. Adding support for callouts in JIT, and fixing some issues revealed
820 during this work. Namely:
822 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
824 (b) Minimum length was not checked before the matching is started.
826 11. The value of capture_last that is passed to callouts was incorrect in some
827 cases when there was a capture on one path that was subsequently abandoned
828 after a backtrack. Also, the capture_last value is now reset after a
829 recursion, since all captures are also reset in this case.
831 12. The interpreter no longer returns the "too many substrings" error in the
832 case when an overflowing capture is in a branch that is subsequently
833 abandoned after a backtrack.
835 13. In the pathological case when an offset vector of size 2 is used, pcretest
836 now prints out the matched string after a yield of 0 or 1.
838 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
839 Only supported by the JIT compiler at the moment.
841 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
843 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
844 offset of the starting point of the matching process, provided the offsets
845 vector is large enough.
847 17. The \A escape now records a lookbehind value of 1, though its execution
848 does not actually inspect the previous character. This is to ensure that,
849 in partial multi-segment matching, at least one character from the old
850 segment is retained when a new segment is processed. Otherwise, if there
851 are no lookbehinds in the pattern, \A might match incorrectly at the start
852 of a new segment.
854 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
856 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
857 gives some modest performance improvement in 8-bit mode.
859 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
860 be expressed in certain programming languages using Universal Character
861 Names.
863 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
864 which points out that "non characters" are not "characters that may not
865 appear in Unicode strings" but rather "characters that are reserved for
866 internal use and have only local meaning".
868 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
869 there was a conditional group that depended on an assertion, if the
870 assertion was false, the callout that immediately followed the alternation
871 in the condition was skipped when pcre_exec() was used for matching.
873 23. Allow an explicit callout to be inserted before an assertion that is the
874 condition for a conditional group, for compatibility with automatic
875 callouts, which always insert a callout at this point.
877 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
878 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
880 25. (*PRUNE) is now supported by the JIT compiler.
882 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
884 27. Fix the case where there are two or more SKIPs with arguments that may be
885 ignored.
887 28. (*SKIP) is now supported by the JIT compiler.
889 29. (*THEN) is now supported by the JIT compiler.
891 30. Update RunTest with additional test selector options.
893 31. The way PCRE handles backtracking verbs has been changed in two ways.
895 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
896 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
897 backtracking. In some cases this makes it more Perl-compatible, but Perl's
898 rather obscure rules do not always do the same thing.
900 (2) Previously, backtracking verbs were confined within assertions. This is
901 no longer the case for positive assertions, except for (*ACCEPT). Again,
902 this sometimes improves Perl compatibility, and sometimes does not.
904 32. A number of tests that were in test 2 because Perl did things differently
905 have been moved to test 1, because either Perl or PCRE has changed, and
906 these tests are now compatible.
908 32. Backtracking control verbs are now handled in the same way in JIT and
909 interpreter.
911 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
912 contained a forward subroutine reference caused a compile error.
914 34. Auto-detect and optimize limited repetitions in JIT.
916 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
917 blocking (*UTF) etc.
919 36. In the interpreter, maximizing pattern repetitions for characters and
920 character types now use tail recursion, which reduces stack usage.
922 37. The value of the max lookbehind was not correctly preserved if a compiled
923 and saved regex was reloaded on a host of different endianness.
925 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
926 of the compiled pattern block, expand the flags field from 16 to 32 bits
927 because it was almost full.
929 39. Try madvise first before posix_madvise.
931 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
932 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
933 original change locked out all matches of empty strings. This has been
934 changed so that one match of an empty string per line is recognized.
935 Subsequent searches on the same line (for colouring or for --only-matching,
936 for example) do not recognize empty strings.
938 41. Applied a user patch to fix a number of spelling mistakes in comments.
940 42. Data lines longer than 65536 caused pcretest to crash.
942 43. Clarified the data type for length and startoffset arguments for pcre_exec
943 and pcre_dfa_exec in the function-specific man pages, where they were
944 explicitly stated to be in bytes, never having been updated. I also added
945 some clarification to the pcreapi man page.
947 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
948 a segmentation fault.
951 Version 8.32 30-November-2012
952 -----------------------------
954 1. Improved JIT compiler optimizations for first character search and single
955 character iterators.
957 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
958 Patch by Daniel Richard G.
960 3. Single character iterator optimizations in the JIT compiler.
962 4. Improved JIT compiler optimizations for character ranges.
964 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
965 Reported by Giuseppe D'Angelo.
967 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
968 of a line, was being set incorrectly in cases where .* appeared inside
969 atomic brackets at the start of a pattern, or where there was a subsequent
970 *PRUNE or *SKIP.
972 7. Improved instruction cache flush for POWER/PowerPC.
973 Patch by Daniel Richard G.
975 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
976 grep:
978 (a) There is now no limit to the number of patterns to be matched.
980 (b) An error is given if a pattern is too long.
982 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
983 are now supported.
985 (d) --exclude-from and --include-from (multiple use) have been added.
987 (e) Exclusions and inclusions now apply to all files and directories, not
988 just to those obtained from scanning a directory recursively.
990 (f) Multiple uses of -f and --file-list are now supported.
992 (g) In a Windows environment, the default for -d has been changed from
993 "read" (the GNU grep default) to "skip", because otherwise the presence
994 of a directory in the file list provokes an error.
996 (h) The documentation has been revised and clarified in places.
998 9. Improve the matching speed of capturing brackets.
1000 10. Changed the meaning of \X so that it now matches a Unicode extended
1001 grapheme cluster.
1003 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1004 out POSIX threads when JIT support is configured.
1006 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1008 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1009 structure could go wrong in environments where size_t is not the same size
1010 as int.
1012 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1013 set.
1015 15. The EBCDIC support had decayed; later updates to the code had included
1016 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1017 general tidy up of EBCDIC-related issues, and the documentation was also
1018 not quite right. There is now a test that can be run on ASCII systems to
1019 check some of the EBCDIC-related things (but is it not a full test).
1021 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1022 in a small tidy to the code.
1024 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1026 18. If the --only-matching (-o) option in pcregrep is specified multiple
1027 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1028 substrings matched by the 1st and 2nd capturing parentheses. A separating
1029 string can be specified by --om-separator (default empty).
1031 19. Improving the first n character searches.
1033 20. Turn case lists for horizontal and vertical white space into macros so that
1034 they are defined only once.
1036 21. This set of changes together give more compatible Unicode case-folding
1037 behaviour for characters that have more than one other case when UCP
1038 support is available.
1040 (a) The Unicode property table now has offsets into a new table of sets of
1041 three or more characters that are case-equivalent. The MultiStage2.py
1042 script that generates these tables (the pcre_ucd.c file) now scans
1043 CaseFolding.txt instead of UnicodeData.txt for character case
1044 information.
1046 (b) The code for adding characters or ranges of characters to a character
1047 class has been abstracted into a generalized function that also handles
1048 case-independence. In UTF-mode with UCP support, this uses the new data
1049 to handle characters with more than one other case.
1051 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1052 whose other case is greater than 256 are now correctly matched
1053 caselessly. Previously, the high codepoint matched the low one, but not
1054 vice versa.
1056 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1057 of the new class addition function, using character lists defined as
1058 macros alongside the case definitions of 20 above.
1060 (e) Caseless back references now work with characters that have more than
1061 one other case.
1063 (f) General caseless matching of characters with more than one other case
1064 is supported.
1066 22. Unicode character properties were updated from Unicode 6.2.0
1068 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1070 24. Add support for 32-bit character strings, and UTF-32
1072 25. Major JIT compiler update (code refactoring and bugfixing).
1073 Experimental Sparc 32 support is added.
1075 26. Applied a modified version of Daniel Richard G's patch to create
1076 pcre.h.generic and config.h.generic by "make" instead of in the
1077 PrepareRelease script.
1079 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1080 pcre_compile.c when checking for a zero character.
1082 28. Introducing a native interface for JIT. Through this interface, the compiled
1083 machine code can be directly executed. The purpose of this interface is to
1084 provide fast pattern matching, so several sanity checks are not performed.
1085 However, feature tests are still performed. The new interface provides
1086 1.4x speedup compared to the old one.
1088 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1089 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1090 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1092 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1093 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1094 the "old" RFC 2279). Instead, it ended up passing a negative length to
1095 pcre_exec().
1097 31. Add support for GCC's visibility feature to hide internal functions.
1099 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1100 "unknown -C option" after outputting 0 or 1.
1102 33. There is now support for generating a code coverage report for the test
1103 suite in environments where gcc is the compiler and lcov is installed. This
1104 is mainly for the benefit of the developers.
1106 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1107 unaddressable using valgrind annotations, allowing valgrind to detect
1108 invalid memory accesses. This is mainly for the benefit of the developers.
1110 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1112 26. Give configure error if --enable-cpp but no C++ compiler found.
1115 Version 8.31 06-July-2012
1116 -------------------------
1118 1. Fixing a wrong JIT test case and some compiler warnings.
1120 2. Removed a bashism from the RunTest script.
1122 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1123 to unsigned type, result still unsigned" that was given by an MS compiler
1124 on encountering the code "-sizeof(xxx)".
1126 4. Partial matching support is added to the JIT compiler.
1128 5. Fixed several bugs concerned with partial matching of items that consist
1129 of more than one character:
1131 (a) /^(..)\1/ did not partially match "aba" because checking references was
1132 done on an "all or nothing" basis. This also applied to repeated
1133 references.
1135 (b) \R did not give a hard partial match if \r was found at the end of the
1136 subject.
1138 (c) \X did not give a hard partial match after matching one or more
1139 characters at the end of the subject.
1141 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1142 a partial match for the string "\r".
1144 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1145 a partial match for a CR character at the end of the subject string.
1147 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1148 running pcretest, the text "(JIT)" added to the output whenever JIT is
1149 actually used to run the match.
1151 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1152 or /S+[+] with a digit between 1 and 7.
1154 8. OP_NOT now supports any UTF character not just single-byte ones.
1156 9. (*MARK) control verb is now supported by the JIT compiler.
1158 10. The command "./RunTest list" lists the available tests without actually
1159 running any of them. (Because I keep forgetting what they all are.)
1163 12. Applied a (slightly modified) user-supplied patch that improves performance
1164 when the heap is used for recursion (compiled with --disable-stack-for-
1165 recursion). Instead of malloc and free for each heap frame each time a
1166 logical recursion happens, frames are retained on a chain and re-used where
1167 possible. This sometimes gives as much as 30% improvement.
1169 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1170 call.
1172 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1174 15. It is now possible to link pcretest with libedit as an alternative to
1175 libreadline.
1177 16. (*COMMIT) control verb is now supported by the JIT compiler.
1179 17. The Unicode data tables have been updated to Unicode 6.1.0.
1181 18. Added --file-list option to pcregrep.
1183 19. Added binary file support to pcregrep, including the -a, --binary-files,
1184 -I, and --text options.
1186 20. The madvise function is renamed for posix_madvise for QNX compatibility
1187 reasons. Fixed by Giuseppe D'Angelo.
1189 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1191 22. Changed the option for creating symbolic links for 16-bit man pages from
1192 -s to -sf so that re-installing does not cause issues.
1194 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1196 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1197 matches in certain environments (the workspace was not being correctly
1198 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1199 some of the workspace data at the beginning of a restart.
1201 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1202 was not doing so when it should - probably a typo introduced by SVN 528
1203 (change 8.10/14).
1205 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1206 \w+ when the character tables indicated that \x{c4} was a word character.
1207 There were several related cases, all because the tests for doing a table
1208 lookup were testing for characters less than 127 instead of 255.
1210 27. If a pattern contains capturing parentheses that are not used in a match,
1211 their slots in the ovector are set to -1. For those that are higher than
1212 any matched groups, this happens at the end of processing. In the case when
1213 there were back references that the ovector was too small to contain
1214 (causing temporary malloc'd memory to be used during matching), and the
1215 highest capturing number was not used, memory off the end of the ovector
1216 was incorrectly being set to -1. (It was using the size of the temporary
1217 memory instead of the true size.)
1219 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1220 ovector size, it uses memory at the end of the block that it has got.
1222 29. Check for an overlong MARK name and give an error at compile time. The
1223 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1225 30. JIT compiler update.
1227 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1228 Rill for the patch.
1230 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1232 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1234 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1235 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1236 was enabled.
1238 35. Improve JIT code generation for greedy plus quantifier.
1240 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1241 1 to "aa" instead of to an empty string. The bug affected repeated groups
1242 that could potentially match an empty string.
1244 37. Optimizing single character iterators in JIT.
1246 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1247 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1248 codepoints that are too big for the mode are faulted, and in a UTF mode,
1249 disallowed codepoints are also faulted.
1251 39. If PCRE was compiled with UTF support, in three places in the DFA
1252 matcher there was code that should only have been obeyed in UTF mode, but
1253 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1254 processing when bytes with values greater than 127 were present. In 16-bit
1255 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1256 both cases the values are those that cannot be the first data item in a UTF
1257 character. The three items that might have provoked this were recursions,
1258 possessively repeated groups, and atomic groups.
1260 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1261 and pcregrep, because some OS require shared objects to be explicitly
1262 passed to ld, causing the link step to fail if they are not.
1264 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1265 mode, patterns that started with \h* or \R* might be incorrectly matched.
1268 Version 8.30 04-February-2012
1269 -----------------------------
1271 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1272 name is defined in ctype.h.
1274 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1275 only in quite long subpatterns.
1277 3. Removed the function pcre_info(), which has been obsolete and deprecated
1278 since it was replaced by pcre_fullinfo() in February 2000.
1280 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1281 match a (*MARK), and the match failed at the start of the subject, a
1282 reference to memory before the start of the subject could occur. This bug
1283 was introduced by fix 17 of release 8.21.
1285 5. A reference to an unset group with zero minimum repetition was giving
1286 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1287 /(another)?(\1?)test/ matched against "hello world test". This bug was
1288 introduced in release 8.13.
1290 6. Add support for 16-bit character strings (a large amount of work involving
1291 many changes and refactorings).
1293 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1294 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1295 from a file.
1297 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1298 rounding is not applied in this particular case).
1300 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1301 if they appear, or are escaped, in patterns.
1303 10. Get rid of a number of -Wunused-but-set-variable warnings.
1305 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1306 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1307 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1308 also returns the mark "x". This bug applied to capturing parentheses,
1309 non-capturing parentheses, and atomic parentheses. It also applied to some
1310 assertions.
1312 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1313 information out of configure.ac instead of relying on pcre.h.generic, which
1314 is not stored in the repository.
1316 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1317 -lreadline.
1319 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1321 15. Applied Graycode's patch to put the top-level frame on the stack rather
1322 than the heap when not using the stack for recursion. This gives a
1323 performance improvement in many cases when recursion is not deep.
1325 16. Experimental code added to "pcretest -C" to output the stack frame size.
1328 Version 8.21 12-Dec-2011
1329 ------------------------
1331 1. Updating the JIT compiler.
1333 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1334 are added as well.
1336 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1337 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1338 calling _pcre_jit_exec. Some extra comments are added.
1340 4. (*MARK) settings inside atomic groups that do not contain any capturing
1341 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1342 was introduced by change 18 for 8.20.
1344 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1345 ECMA-262 standard.
1347 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1348 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1349 This bug was probably introduced by change 9 of 8.13.
1351 7. While fixing 6 above, I noticed that a number of other items were being
1352 incorrectly rejected as "not fixed length". This arose partly because newer
1353 opcodes had not been added to the fixed-length checking code. I have (a)
1354 corrected the bug and added tests for these items, and (b) arranged for an
1355 error to occur if an unknown opcode is encountered while checking for fixed
1356 length instead of just assuming "not fixed length". The items that were
1357 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1358 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1359 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1361 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1362 being incorrectly compiled and would have given unpredicatble results.
1364 9. A possessively repeated subpattern with minimum repeat count greater than
1365 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1366 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1367 the first (A) could occur when it should not.
1369 10. Add a cast and remove a redundant test from the code.
1371 11. JIT should use pcre_malloc/pcre_free for allocation.
1373 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1374 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1375 is anything other than /usr, -L is still shown).
1377 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1379 14. Perl does not support \N without a following name in a [] class; PCRE now
1380 also gives an error.
1382 15. If a forward reference was repeated with an upper limit of around 2000,
1383 it caused the error "internal error: overran compiling workspace". The
1384 maximum number of forward references (including repeats) was limited by the
1385 internal workspace, and dependent on the LINK_SIZE. The code has been
1386 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1387 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1388 of around 200,000 forward references. While doing this, I also speeded up
1389 the filling in of repeated forward references.
1391 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1392 incorrectly expecting the subject to contain another "a" after the start.
1394 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1395 in the match, the SKIP should be ignored. This was not happening; instead
1396 the SKIP was being treated as NOMATCH. For patterns such as
1397 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1398 tested.
1400 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1401 now much more compatible with Perl, in particular in cases where the result
1402 is a non-match for a non-anchored pattern. For example, if
1403 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1404 "m", where previously it did not return a name. A side effect of this
1405 change is that for partial matches, the last encountered mark name is
1406 returned, as for non matches. A number of tests that were previously not
1407 Perl-compatible have been moved into the Perl-compatible test files. The
1408 refactoring has had the pleasing side effect of removing one argument from
1409 the match() function, thus reducing its stack requirements.
1411 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1412 subsequent uses of /S (without +) incorrectly behaved like /S+.
1414 21. Retrieve executable code size support for the JIT compiler and fixing
1415 some warnings.
1417 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1418 not work when the shorter character appeared right at the end of the
1419 subject string.
1421 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1422 systems.
1424 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1425 output it when the /M option is used in pcretest.
1427 25. The CheckMan script was not being included in the distribution. Also, added
1428 an explicit "perl" to run Perl scripts from the PrepareRelease script
1429 because this is reportedly needed in Windows.
1431 26. If study data was being save in a file and studying had not found a set of
1432 "starts with" bytes for the pattern, the data written to the file (though
1433 never used) was taken from uninitialized memory and so caused valgrind to
1434 complain.
1436 27. Updated RunTest.bat as provided by Sheri Pierce.
1438 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1440 29. Computation of memory usage for the table of capturing group names was
1441 giving an unnecessarily large value.
1444 Version 8.20 21-Oct-2011
1445 ------------------------
1447 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1448 a POSIX class. After further experiments with Perl, which convinced me that
1449 Perl has bugs and confusions, a closing square bracket is no longer allowed
1450 in a POSIX name. This bug also affected patterns with classes that started
1451 with full stops.
1453 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1454 captured substring, but while checking the failing first alternative,
1455 substring 1 is temporarily captured. If the output vector supplied to
1456 pcre_exec() was not big enough for this capture, the yield of the function
1457 was still zero ("insufficient space for captured substrings"). This cannot
1458 be totally fixed without adding another stack variable, which seems a lot
1459 of expense for a edge case. However, I have improved the situation in cases
1460 such as /(a)(b)x|abc/ matched against "abc", where the return code
1461 indicates that fewer than the maximum number of slots in the ovector have
1462 been set.
1464 3. Related to (2) above: when there are more back references in a pattern than
1465 slots in the output vector, pcre_exec() uses temporary memory during
1466 matching, and copies in the captures as far as possible afterwards. It was
1467 using the entire output vector, but this conflicts with the specification
1468 that only 2/3 is used for passing back captured substrings. Now it uses
1469 only the first 2/3, for compatibility. This is, of course, another edge
1470 case.
1472 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1473 main code base, and can be used by building with --enable-jit. When this is
1474 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1475 runtime --no-jit option is given.
1477 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1478 ovector, the return from the function was zero, implying that there were
1479 other matches that did not fit. The correct "exactly full" value is now
1480 returned.
1482 6. If a subpattern that was called recursively or as a subroutine contained
1483 (*PRUNE) or any other control that caused it to give a non-standard return,
1484 invalid errors such as "Error -26 (nested recursion at the same subject
1485 position)" or even infinite loops could occur.
1487 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1488 computing the minimum length on reaching *ACCEPT, and so ended up with the
1489 wrong value of 1 rather than 0. Further investigation indicates that
1490 computing a minimum subject length in the presence of *ACCEPT is difficult
1491 (think back references, subroutine calls), and so I have changed the code
1492 so that no minimum is registered for a pattern that contains *ACCEPT.
1494 8. If (*THEN) was present in the first (true) branch of a conditional group,
1495 it was not handled as intended. [But see 16 below.]
1497 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1498 Sheri Pierce.
1500 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1501 the first byte in a match must be "a".
1503 11. Change 17 for 8.13 increased the recursion depth for patterns like
1504 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1505 pattern contains any instances of (*THEN). If it does not, the old
1506 optimizations are restored. It would be nice to do this on a per-group
1507 basis, but at the moment that is not feasible.
1509 12. In some environments, the output of pcretest -C is CRLF terminated. This
1510 broke RunTest's code that checks for the link size. A single white space
1511 character after the value is now allowed for.
1513 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1514 For "fr", it uses the Windows-specific input and output files.
1516 14. If (*THEN) appeared in a group that was called recursively or as a
1517 subroutine, it did not work as intended. [But see next item.]
1519 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1520 pattern fragments (but not containing any | characters). If A and B are
1521 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1522 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1523 D. In other words, Perl considers parentheses that do not contain any |
1524 characters to be part of a surrounding alternative, whereas PCRE was
1525 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1526 differently. PCRE now behaves in the same way as Perl, except in the case
1527 of subroutine/recursion calls such as (?1) which have in any case always
1528 been different (but PCRE had them first :-).
1530 16. Related to 15 above: Perl does not treat the | in a conditional group as
1531 creating alternatives. Such a group is treated in the same way as an
1532 ordinary group without any | characters when processing (*THEN). PCRE has
1533 been changed to match Perl's behaviour.
1535 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1536 RunGrepTest script failed.
1538 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1539 inevitable for groups that contain captures, but it can lead to a lot of
1540 stack use in large patterns. The old behaviour has been restored for atomic
1541 groups that do not contain any capturing parentheses.
1543 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1544 suppress the check for a minimum subject length at run time. (If it was
1545 given to pcre_exec() or pcre_dfa_exec() it did work.)
1547 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1548 fail to work when decoding hex characters in data strings in EBCDIC
1549 environments.
1551 21. It appears that in at least one Mac OS environment, the isxdigit() function
1552 is implemented as a macro that evaluates to its argument more than once,
1553 contravening the C 90 Standard (I haven't checked a later standard). There
1554 was an instance in pcretest which caused it to go wrong when processing
1555 \x{...} escapes in subject strings. The has been rewritten to avoid using
1556 things like p++ in the argument of isxdigit().
1559 Version 8.13 16-Aug-2011
1560 ------------------------
1562 1. The Unicode data tables have been updated to Unicode 6.0.0.
1564 2. Two minor typos in pcre_internal.h have been fixed.
1566 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1567 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1568 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1570 4. There were a number of related bugs in the code for matching backrefences
1571 caselessly in UTF-8 mode when codes for the characters concerned were
1572 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1573 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1574 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1575 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1576 2-byte code at the end of the subject (it thought there wasn't enough data
1577 left).
1579 5. Comprehensive information about what went wrong is now returned by
1580 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1581 as the output vector has at least 2 elements. The offset of the start of
1582 the failing character and a reason code are placed in the vector.
1584 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1585 now returned is for the first byte of the failing character, instead of the
1586 last byte inspected. This is an incompatible change, but I hope it is small
1587 enough not to be a problem. It makes the returned offset consistent with
1588 pcre_exec() and pcre_dfa_exec().
1590 7. pcretest now gives a text phrase as well as the error number when
1591 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1592 failure, the offset and reason code are output.
1594 8. When \R was used with a maximizing quantifier it failed to skip backwards
1595 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1596 back over a single character (\n). This seems wrong (because it treated the
1597 two characters as a single entity when going forwards), conflicts with the
1598 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1599 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1600 has been changed.
1602 9. Some internal refactoring has changed the processing so that the handling
1603 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1604 time (the PCRE_DOTALL option was changed this way some time ago: version
1605 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1606 which was always a bit of a fudge. It also means that there is one less
1607 argument for the match() function, which reduces its stack requirements
1608 slightly. This change also fixes an incompatibility with Perl: the pattern
1609 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1611 10. More internal refactoring has drastically reduced the number of recursive
1612 calls to match() for possessively repeated groups such as (abc)++ when
1613 using pcre_exec().
1615 11. While implementing 10, a number of bugs in the handling of groups were
1616 discovered and fixed:
1618 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1619 (a|)*(?1) gave a compile-time internal error.
1620 ((a|)+)+ did not notice that the outer group could match an empty string.
1621 (^a|^)+ was not marked as anchored.
1622 (.*a|.*)+ was not marked as matching at start or after a newline.
1624 12. Yet more internal refactoring has removed another argument from the match()
1625 function. Special calls to this function are now indicated by setting a
1626 value in a variable in the "match data" data block.
1628 13. Be more explicit in pcre_study() instead of relying on "default" for
1629 opcodes that mean there is no starting character; this means that when new
1630 ones are added and accidentally left out of pcre_study(), testing should
1631 pick them up.
1633 14. The -s option of pcretest has been documented for ages as being an old
1634 synonym of -m (show memory usage). I have changed it to mean "force study
1635 for every regex", that is, assume /S for every regex. This is similar to -i
1636 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1637 using it. It makes it easier to run collections of tests with and without
1638 study enabled, and thereby test pcre_study() more easily. All the standard
1639 tests are now run with and without -s (but some patterns can be marked as
1640 "never study" - see 20 below).
1642 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1643 restoration of the capturing data to the outer values was not happening
1644 correctly.
1646 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1647 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1648 pattern had matched an empty string, and so incorrectly returned a no
1649 match.
1651 17. There was optimizing code for the last branch of non-capturing parentheses,
1652 and also for the obeyed branch of a conditional subexpression, which used
1653 tail recursion to cut down on stack usage. Unfortunately, now that there is
1654 the possibility of (*THEN) occurring in these branches, tail recursion is
1655 no longer possible because the return has to be checked for (*THEN). These
1656 two optimizations have therefore been removed. [But see 8.20/11 above.]
1658 18. If a pattern containing \R was studied, it was assumed that \R always
1659 matched two bytes, thus causing the minimum subject length to be
1660 incorrectly computed because \R can also match just one byte.
1662 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1663 was incorrectly computed.
1665 20. If /S is present twice on a test pattern in pcretest input, it now
1666 *disables* studying, thereby overriding the use of -s on the command line
1667 (see 14 above). This is necessary for one or two tests to keep the output
1668 identical in both cases.
1670 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1671 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1673 22. When an atomic group that contained a capturing parenthesis was
1674 successfully matched, but the branch in which it appeared failed, the
1675 capturing was not being forgotten if a higher numbered group was later
1676 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1677 group 1 to "a", when in fact it should be unset. This applied to multi-
1678 branched capturing and non-capturing groups, repeated or not, and also to
1679 positive assertions (capturing in negative assertions does not happen
1680 in PCRE) and also to nested atomic groups.
1682 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1683 subject after a captured substring, to make it easier to tell which of a
1684 number of identical substrings has been captured.
1686 24. The way atomic groups are processed by pcre_exec() has been changed so that
1687 if they are repeated, backtracking one repetition now resets captured
1688 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1689 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1690 "aaa". Previously, it would have been "a". As part of this code
1691 refactoring, the way recursive calls are handled has also been changed.
1693 25. If an assertion condition captured any substrings, they were not passed
1694 back unless some other capturing happened later. For example, if
1695 (?(?=(a))a) was matched against "a", no capturing was returned.
1697 26. When studying a pattern that contained subroutine calls or assertions,
1698 the code for finding the minimum length of a possible match was handling
1699 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1700 group 1 called group 2 while simultaneously a separate group 2 called group
1701 1). A stack overflow occurred in this case. I have fixed this by limiting
1702 the recursion depth to 10.
1704 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1705 Fortmann. This supports explicit test numbers on the command line, and has
1706 argument validation and error reporting.
1708 28. An instance of \X with an unlimited repeat could fail if at any point the
1709 first character it looked at was a mark character.
1711 29. Some minor code refactoring concerning Unicode properties and scripts
1712 should reduce the stack requirement of match() slightly.
1714 30. Added the '=' option to pcretest to check the setting of unused capturing
1715 slots at the end of the pattern, which are documented as being -1, but are
1716 not included in the return count.
1718 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1719 compiled something random. Now it gives a compile-time error (as does
1720 Perl).
1722 32. A *MARK encountered during the processing of a positive assertion is now
1723 recorded and passed back (compatible with Perl).
1725 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1726 had alternative anchored branches, the search for a second match in a line
1727 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1728 matched the line "0102" twice. The same bug affected patterns that started
1729 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1730 twice.
1732 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1733 does, and because of capturing effects, quantifying parenthesized
1734 assertions may at times be useful. Quantifiers are now allowed for
1735 parenthesized assertions.
1737 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1739 36. \g was being checked for fancy things in a character class, when it should
1740 just be a literal "g".
1742 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1743 appearance of a nested POSIX class supersedes an apparent external class.
1744 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1745 unescaped square brackets may also appear as part of class names. For
1746 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1747 more like Perl. (But see 8.20/1 above.)
1749 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1750 was because it thought it was \N{name}, which is not supported).
1752 39. Add minix to OS list not supporting the -S option in pcretest.
1754 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1755 cannot analyze patterns in sufficient detail to catch mutual recursions
1756 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1757 subgroup is called recursively as a subpattern for a second time at the
1758 same position in the subject string. In previous releases this might have
1759 been caught by the recursion limit, or it might have run out of stack.
1761 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1762 happen only once. PCRE was, however incorrectly giving a compile time error
1763 "recursive call could loop indefinitely" because it cannot analyze the
1764 pattern in sufficient detail. The compile time test no longer happens when
1765 PCRE is compiling a conditional subpattern, but actual runaway loops are
1766 now caught at runtime (see 40 above).
1768 42. It seems that Perl allows any characters other than a closing parenthesis
1769 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1770 has been changed to be the same.
1772 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1773 as not to get warnings when autogen.sh is called. Also changed
1774 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1776 44. To help people who use pcregrep to scan files containing exceedingly long
1777 lines, the following changes have been made:
1779 (a) The default value of the buffer size parameter has been increased from
1780 8K to 20K. (The actual buffer used is three times this size.)
1782 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1783 PCRE is built.
1785 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1786 to be set at run time.
1788 (d) Numerical values in pcregrep options can be followed by K or M, for
1789 example --buffer-size=50K.
1791 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1792 given and the return code is set to 2.
1794 45. Add a pointer to the latest mark to the callout data block.
1796 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1797 partial match of an empty string instead of no match. This was specific to
1798 the use of ".".
1800 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1801 complete match instead of a partial match. This bug was dependent on both
1802 the PCRE_UTF8 and PCRE_DOTALL options being set.
1804 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1805 starting byte set, because \b was not being ignored.
1808 Version 8.12 15-Jan-2011
1809 ------------------------
1811 1. Fixed some typos in the markup of the man pages, and wrote a script that
1812 checks for such things as part of the documentation building process.
1814 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1815 --match-limit and --recursion-limit options (added for 8.11). In
1816 particular, this made one of the standard tests fail. (The integer value
1817 went into the wrong half of a long int.)
1819 3. If the --colour option was given to pcregrep with -v (invert match), it
1820 did strange things, either producing crazy output, or crashing. It should,
1821 of course, ignore a request for colour when reporting lines that do not
1822 match.
1824 4. Another pcregrep bug caused similar problems if --colour was specified with
1825 -M (multiline) and the pattern match finished with a line ending.
1827 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1828 matched in multiline mode, the following line was shown as part of the
1829 match. This seems wrong, so I have changed it.
1831 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1832 the check for further matches in the same line (so they could be coloured)
1833 to overrun the end of the current line. If another match was found, it was
1834 incorrectly shown (and then shown again when found in the next line).
1836 7. If pcregrep was compiled under Windows, there was a reference to the
1837 function pcregrep_exit() before it was defined. I am assuming this was
1838 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1839 reported by a user. I've moved the definition above the reference.
1842 Version 8.11 10-Dec-2010
1843 ------------------------
1845 1. (*THEN) was not working properly if there were untried alternatives prior
1846 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1847 backtracked to try for "b" instead of moving to the next alternative branch
1848 at the same level (in this case, to look for "c"). The Perl documentation
1849 is clear that when (*THEN) is backtracked onto, it goes to the "next
1850 alternative in the innermost enclosing group".
1852 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1853 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1854 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1855 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1856 (*THEN).
1858 3. If \s appeared in a character class, it removed the VT character from
1859 the class, even if it had been included by some previous item, for example
1860 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1861 of \s, but is part of the POSIX "space" class.)
1863 4. A partial match never returns an empty string (because you can always
1864 match an empty string at the end of the subject); however the checking for
1865 an empty string was starting at the "start of match" point. This has been
1866 changed to the "earliest inspected character" point, because the returned
1867 data for a partial match starts at this character. This means that, for
1868 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1869 (previously it gave "no match").
1871 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1872 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1873 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1874 has an implication that the given string is incomplete (because a partial
1875 match is preferred over a full match). For this reason, these items now
1876 give a partial match in this situation. [Aside: previously, the one case
1877 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1878 match rather than a full match, which was wrong by the old rules, but is
1879 now correct.]
1881 6. There was a bug in the handling of #-introduced comments, recognized when
1882 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1883 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1884 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1885 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1886 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1887 places in pcre_compile().
1889 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1890 comments when looking ahead for named forward references to subpatterns,
1891 the only newline sequence it recognized was NL. It now handles newlines
1892 according to the set newline convention.
1894 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1895 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1896 cater for a lack of strerror(). These oversights have been fixed.
1898 9. Added --match-limit and --recursion-limit to pcregrep.
1900 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1902 11. When the -o option was used, pcregrep was setting a return code of 1, even
1903 when matches were found, and --line-buffered was not being honoured.
1905 12. Added an optional parentheses number to the -o and --only-matching options
1906 of pcregrep.
1908 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1909 can match an empty string. The code to do it in pcretest and pcredemo
1910 needed fixing:
1912 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1913 only one byte after an empty string match just before CRLF (this case
1914 just got forgotten; "any" and "anycrlf" were OK).
1916 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1917 mode when an empty string match preceded an ASCII character followed by
1918 a non-ASCII character. (The code for advancing by one character rather
1919 than one byte was nonsense.)
1921 (c) The pcredemo.c sample program did not have any code at all to handle
1922 the cases when CRLF is a valid newline sequence.
1924 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1925 as a starting offset was within the subject string. There is now a new
1926 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1927 negative or greater than the length of the string. In order to test this,
1928 pcretest is extended to allow the setting of negative starting offsets.
1930 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1931 starting offset points to the beginning of a UTF-8 character was
1932 unnecessarily clumsy. I tidied it up.
1934 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1935 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1937 17. Nobody had reported that the --include_dir option, which was added in
1938 release 7.7 should have been called --include-dir (hyphen, not underscore)
1939 for compatibility with GNU grep. I have changed it to --include-dir, but
1940 left --include_dir as an undocumented synonym, and the same for
1941 --exclude-dir, though that is not available in GNU grep, at least as of
1942 release 2.5.4.
1944 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1945 characters from a string of bytes) have been redefined so as not to use
1946 loops, in order to improve performance in some environments. At the same
1947 time, I abstracted some of the common code into auxiliary macros to save
1948 repetition (this should not affect the compiled code).
1950 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1951 compile-time error is now given if \c is not followed by an ASCII
1952 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1953 different, and any byte value is allowed.)
1955 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1956 START_OPTIMIZE option, which is now allowed at compile time - but just
1957 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1958 to pcregrep and other applications that have no direct access to PCRE
1959 options. The new /Y option in pcretest sets this option when calling
1960 pcre_compile().
1962 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1963 back references. Groups containing recursive back references were forced to
1964 be atomic by that change, but in the case of named groups, the amount of
1965 memory required was incorrectly computed, leading to "Failed: internal
1966 error: code overflow". This has been fixed.
1968 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1969 pcretest.c, to avoid build problems in some Borland environments.
1972 Version 8.10 25-Jun-2010
1973 ------------------------
1975 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1976 THEN.
1978 2. (*ACCEPT) was not working when inside an atomic group.
1980 3. Inside a character class, \B is treated as a literal by default, but
1981 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1982 causes the error). The code is unchanged, but I tidied the documentation.
1984 4. Inside a character class, PCRE always treated \R and \X as literals,
1985 whereas Perl faults them if its -w option is set. I have changed PCRE so
1986 that it faults them when PCRE_EXTRA is set.
1988 5. Added support for \N, which always matches any character other than
1989 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1991 6. When compiling pcregrep with newer versions of gcc which may have
1992 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1993 declared with attribute warn_unused_result" were given. Just casting the
1994 result to (void) does not stop the warnings; a more elaborate fudge is
1995 needed. I've used a macro to implement this.
1997 7. Minor change to pcretest.c to avoid a compiler warning.
1999 8. Added four artifical Unicode properties to help with an option to make
2000 \s etc use properties (see next item). The new properties are: Xan
2001 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2003 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2004 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2005 this option. Modified pcretest to add /W to test this facility. Added
2006 REG_UCP to make it available via the POSIX interface.
2008 10. Added --line-buffered to pcregrep.
2010 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2011 studied, and the match started with a letter with a code point greater than
2012 127 whose first byte was different to the first byte of the other case of
2013 the letter, the other case of this starting letter was not recognized
2014 (#976).
2016 12. If a pattern that was studied started with a repeated Unicode property
2017 test, for example, \p{Nd}+, there was the theoretical possibility of
2018 setting up an incorrect bitmap of starting bytes, but fortunately it could
2019 not have actually happened in practice until change 8 above was made (it
2020 added property types that matched character-matching opcodes).
2022 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2023 possible starting bytes for non-anchored patterns.
2025 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2026 \R, and also a number of cases that involve Unicode properties, both
2027 explicit and implicit when PCRE_UCP is set.
2029 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2030 input, it could crash or give wrong results if characters with values
2031 greater than 0xc0 were present in the subject string. (Detail: it assumed
2032 UTF-8 input when processing these items.)
2034 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2035 size_t is 64-bit (#991).
2037 17. Added a check for running out of memory when PCRE is compiled with
2038 --disable-stack-for-recursion (#990).
2040 18. If the last data line in a file for pcretest does not have a newline on
2041 the end, a newline was missing in the output.
2043 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2044 less than 128) in its various bitmaps. However, there is a facility for
2045 generating tables according to the current locale when PCRE is compiled. It
2046 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2047 characters, are recognized by isspace() and therefore were getting set in
2048 these tables, and indeed these tables seem to approximate to ISO 8859. This
2049 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2050 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2051 which of course cannot start UTF-8 characters. I have changed the code so
2052 that only real ASCII characters (less than 128) and the correct starting
2053 bytes for UTF-8 encodings are set for characters greater than 127 when in
2054 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2055 altogether.)
2057 20. Added the /T option to pcretest so as to be able to run tests with non-
2058 standard character tables, thus making it possible to include the tests
2059 used for 19 above in the standard set of tests.
2061 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2062 reference to a subpattern the other side of a comment that contains an
2063 opening parenthesis caused either an internal compiling error, or a
2064 reference to the wrong subpattern.
2067 Version 8.02 19-Mar-2010
2068 ------------------------
2070 1. The Unicode data tables have been updated to Unicode 5.2.0.
2072 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2073 configured.
2075 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2076 original author of that file, following a query about its status.
2078 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2079 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2081 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2082 quantifier applied to a forward-referencing subroutine call, could compile
2083 incorrect code or give the error "internal error: previously-checked
2084 referenced subpattern not found".
2086 6. Both MS Visual Studio and Symbian OS have problems with initializing
2087 variables to point to external functions. For these systems, therefore,
2088 pcre_malloc etc. are now initialized to local functions that call the
2089 relevant global functions.
2091 7. There were two entries missing in the vectors called coptable and poptable
2092 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2093 I've fixed the data, and added a kludgy way of testing at compile time that
2094 the lengths are correct (equal to the number of opcodes).
2096 8. Following on from 7, I added a similar kludge to check the length of the
2097 eint vector in pcreposix.c.
2099 9. Error texts for pcre_compile() are held as one long string to avoid too
2100 much relocation at load time. To find a text, the string is searched,
2101 counting zeros. There was no check for running off the end of the string,
2102 which could happen if a new error number was added without updating the
2103 string.
2105 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2107 11. \K was not working if it appeared in an atomic group or in a group that
2108 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2109 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2110 the assertion is positive, but not if it is negative.
2112 12. Change 11 fortuitously reduced the size of the stack frame used in the
2113 "match()" function of pcre_exec.c by one pointer. Forthcoming
2114 implementation of support for (*MARK) will need an extra pointer on the
2115 stack; I have reserved it now, so that the stack frame size does not
2116 decrease.
2118 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2119 item in branch that calls a recursion is a subroutine call - as in the
2120 second branch in the above example - was incorrectly given the compile-
2121 time error "recursive call could loop indefinitely" because pcre_compile()
2122 was not correctly checking the subroutine for matching a non-empty string.
2124 14. The checks for overrunning compiling workspace could trigger after an
2125 overrun had occurred. This is a "should never occur" error, but it can be
2126 triggered by pathological patterns such as hundreds of nested parentheses.
2127 The checks now trigger 100 bytes before the end of the workspace.
2129 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2132 Version 8.01 19-Jan-2010
2133 ------------------------
2135 1. If a pattern contained a conditional subpattern with only one branch (in
2136 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2137 computed the wrong minimum data length (which is of course zero for such
2138 subpatterns). This could cause incorrect "no match" results.
2140 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2141 the pattern is reset in the first branch, pcre_compile() failed with
2142 "internal error: code overflow at offset...". This happened only when
2143 the reset was to the original external option setting. (An optimization
2144 abstracts leading options settings into an external setting, which was the
2145 cause of this.)
2147 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2148 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2149 assertion pattern did not match (meaning that the assertion was true), it
2150 was incorrectly treated as false if the SKIP had been reached during the
2151 matching. This also applied to assertions used as conditions.
2153 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2154 assertion subpattern, including such a pattern used as a condition,
2155 unpredictable results occurred, instead of the error return
2158 5. The C++ GlobalReplace function was not working like Perl for the special
2159 situation when an empty string is matched. It now does the fancy magic
2160 stuff that is necessary.
2162 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2163 removed. (These were left over from very, very early versions of PCRE.)
2165 7. Some cosmetic changes to the code to make life easier when compiling it
2166 as part of something else:
2168 (a) Change DEBUG to PCRE_DEBUG.
2170 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2171 called "current" as "current_branch", to prevent a collision with the
2172 Linux macro when compiled as a kernel module.
2174 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2175 prevent a collision with the Linux macro when compiled as a kernel
2176 module.
2178 8. In pcre_compile() there are some checks for integer overflows that used to
2179 cast potentially large values to (double). This has been changed to that
2180 when building, a check for int64_t is made, and if it is found, it is used
2181 instead, thus avoiding the use of floating point arithmetic. (There is no
2182 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2183 double.
2185 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2186 2005 (difference between two addresses compared to an unsigned value).
2188 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2189 custom one, because of the following reported problem in Windows:
2191 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2192 under Win32.
2193 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2194 therefore missing the function definition.
2195 - The compiler thus generates a "C" signature for the test function.
2196 - The linker fails to find the "C" function.
2197 - PCRE fails to configure if asked to do so against libbz2.
2199 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2200 messages were output:
2202 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2203 rerunning libtoolize, to keep the correct libtool macros in-tree.
2204 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2206 I have done both of these things.
2208 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2209 most of the time, it *can* run out if it is given a pattern that contains a
2210 runaway infinite recursion. I updated the discussion in the pcrestack man
2211 page.
2213 13. Now that we have gone to the x.xx style of version numbers, the minor
2214 version may start with zero. Using 08 or 09 is a bad idea because users
2215 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2216 interpreted as invalid octal numbers. I've updated the previous comment in
2217 configure.ac, and also added a check that gives an error if 08 or 09 are
2218 used.
2220 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2221 causing partial matching to fail when the end of the subject matched \W
2222 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2224 15. There were some discrepancies between the declarations in pcre_internal.h
2225 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2226 their definitions. The declarations used "const uschar *" and the
2227 definitions used USPTR. Even though USPTR is normally defined as "const
2228 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2229 reported that: "This difference in casting confuses some C++ compilers, for
2230 example, SunCC recognizes above declarations as different functions and
2231 generates broken code for hbpcre." I have changed the declarations to use
2232 USPTR.
2234 16. GNU libtool is named differently on some systems. The autogen.sh script now
2235 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2236 (FreeBSD).
2238 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2239 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2240 comment: "Figure out how to create a longlong from a string: strtoll and
2241 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2242 instance, but it only takes 2 args instead of 3!"
2244 18. A subtle bug concerned with back references has been fixed by a change of
2245 specification, with a corresponding code fix. A pattern such as
2246 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2247 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2248 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2249 same bug. Such groups have to be quantified to be useful, or contained
2250 inside another quantified group. (If there's no repetition, the reference
2251 can never match.) The problem arises because, having left the group and
2252 moved on to the rest of the pattern, a later failure that backtracks into
2253 the group uses the captured value from the final iteration of the group
2254 rather than the correct earlier one. I have fixed this in PCRE by forcing
2255 any group that contains a reference to itself to be an atomic group; that
2256 is, there cannot be any backtracking into it once it has completed. This is
2257 similar to recursive and subroutine calls.
2260 Version 8.00 19-Oct-09
2261 ----------------------
2263 1. The table for translating pcre_compile() error codes into POSIX error codes
2264 was out-of-date, and there was no check on the pcre_compile() error code
2265 being within the table. This could lead to an OK return being given in
2266 error.
2268 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2269 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2270 in a Windows environment.
2272 3. The pcregrep --count option prints the count for each file even when it is
2273 zero, as does GNU grep. However, pcregrep was also printing all files when
2274 --files-with-matches was added. Now, when both options are given, it prints
2275 counts only for those files that have at least one match. (GNU grep just
2276 prints the file name in this circumstance, but including the count seems
2277 more useful - otherwise, why use --count?) Also ensured that the
2278 combination -clh just lists non-zero counts, with no names.
2280 4. The long form of the pcregrep -F option was incorrectly implemented as
2281 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2282 but it seems right to fix it, and I didn't think it was worth preserving
2283 the old behaviour.
2285 5. The command line items --regex=pattern and --regexp=pattern were not
2286 recognized by pcregrep, which required --regex pattern or --regexp pattern
2287 (with a space rather than an '='). The man page documented the '=' forms,
2288 which are compatible with GNU grep; these now work.
2290 6. No libpcreposix.pc file was created for pkg-config; there was just
2291 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2293 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2294 when UCP support is not needed, by modifying the Python script that
2295 generates it from Unicode data files. This should not matter if the module
2296 is correctly used as a library, but I received one complaint about 50K of
2297 unwanted data. My guess is that the person linked everything into his
2298 program rather than using a library. Anyway, it does no harm.
2300 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2301 was a minimum greater than 1 for a wide character in a possessive
2302 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2303 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2304 character. Chaos in the form of incorrect output or a compiling loop could
2305 result.
2307 9. The restrictions on what a pattern can contain when partial matching is
2308 requested for pcre_exec() have been removed. All patterns can now be
2309 partially matched by this function. In addition, if there are at least two
2310 slots in the offset vector, the offset of the earliest inspected character
2311 for the match and the offset of the end of the subject are set in them when
2312 PCRE_ERROR_PARTIAL is returned.
2314 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2315 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2316 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2317 and may be more useful for multi-segment matching.
2319 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2320 used to be given if ever the end of the subject was reached; now it is
2321 given only if matching could not proceed because another character was
2322 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2323 string "Z", which now yields "no match" instead of "partial match". In the
2324 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2325 final character ended with (*FAIL).
2327 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2328 if the pattern had a "must contain" character that was already found in the
2329 earlier partial match, unless partial matching was again requested. For
2330 example, with the pattern /dog.(body)?/, the "must contain" character is
2331 "g". If the first part-match was for the string "dog", restarting with
2332 "sbody" failed. This bug has been fixed.
2334 13. The string returned by pcre_dfa_exec() after a partial match has been
2335 changed so that it starts at the first inspected character rather than the
2336 first character of the match. This makes a difference only if the pattern
2337 starts with a lookbehind assertion or \b or \B (\K is not supported by
2338 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2339 matching functions compatible, and I think it's the right thing to do.
2341 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2342 so that the demonstration program is easily available in environments where
2343 PCRE has not been installed from source.
2345 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2346 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2347 library.
2349 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2350 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2351 is not the first non-POSIX option to be added. Clearly some people find
2352 these options useful.
2354 17. If a caller to the POSIX matching function regexec() passes a non-zero
2355 value for nmatch with a NULL value for pmatch, the value of
2356 nmatch is forced to zero.
2358 18. RunGrepTest did not have a test for the availability of the -u option of
2359 the diff command, as RunTest does. It now checks in the same way as
2360 RunTest, and also checks for the -b option.
2362 19. If an odd number of negated classes containing just a single character
2363 interposed, within parentheses, between a forward reference to a named
2364 subpattern and the definition of the subpattern, compilation crashed with
2365 an internal error, complaining that it could not find the referenced
2366 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2367 [The bug was that it was starting one character too far in when skipping
2368 over the character class, thus treating the ] as data rather than
2369 terminating the class. This meant it could skip too much.]
2371 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2372 /g option in pcretest when the pattern contains \K, which makes it possible
2373 to have an empty string match not at the start, even when the pattern is
2374 anchored. Updated pcretest and pcredemo to use this option.
2376 21. If the maximum number of capturing subpatterns in a recursion was greater
2377 than the maximum at the outer level, the higher number was returned, but
2378 with unset values at the outer level. The correct (outer level) value is
2379 now given.
2381 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2382 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2383 make it do so. The string so far is captured, making this feature
2384 compatible with Perl.
2386 23. The tests have been re-organized, adding tests 11 and 12, to make it
2387 possible to check the Perl 5.10 features against Perl 5.10.
2389 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2390 pattern matches a fixed length string. PCRE did not allow this; now it
2391 does. Neither allows recursion.
2393 25. I finally figured out how to implement a request to provide the minimum
2394 length of subject string that was needed in order to match a given pattern.
2395 (It was back references and recursion that I had previously got hung up
2396 on.) This code has now been added to pcre_study(); it finds a lower bound
2397 to the length of subject needed. It is not necessarily the greatest lower
2398 bound, but using it to avoid searching strings that are too short does give
2399 some useful speed-ups. The value is available to calling programs via
2400 pcre_fullinfo().
2402 26. While implementing 25, I discovered to my embarrassment that pcretest had
2403 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2404 study optimizations had never been tested with that matching function.
2405 Oops. What is worse, even when it was passed study data, there was a bug in
2406 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2407 were also very few tests of studied patterns with pcre_dfa_exec().
2409 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2410 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2411 on the other side of the coin, they are no longer allowed to have different
2412 names, because these cannot be distinguished in PCRE, and this has caused
2413 confusion. (This is a difference from Perl.)
2415 28. When duplicate subpattern names are present (necessarily with different
2416 numbers, as required by 27 above), and a test is made by name in a
2417 conditional pattern, either for a subpattern having been matched, or for
2418 recursion in such a pattern, all the associated numbered subpatterns are
2419 tested, and the overall condition is true if the condition is true for any
2420 one of them. This is the way Perl works, and is also more like the way
2421 testing by number works.
2424 Version 7.9 11-Apr-09
2425 ---------------------
2427 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2428 (pcretest), all targets were linked against these libraries. This included
2429 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2430 libraries. This caused unwanted dependencies to be created. This problem
2431 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2432 pcretest is linked with readline.
2434 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2435 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2436 moved outside it again, because FALSE and TRUE are already defined in AIX,
2437 but BOOL is not.
2439 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2440 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2442 4. The pcregrep documentation said spaces were inserted as well as colons (or
2443 hyphens) following file names and line numbers when outputting matching
2444 lines. This is not true; no spaces are inserted. I have also clarified the
2445 wording for the --colour (or --color) option.
2447 5. In pcregrep, when --colour was used with -o, the list of matching strings
2448 was not coloured; this is different to GNU grep, so I have changed it to be
2449 the same.
2451 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2452 each matching line was coloured. Now it goes on to look for further matches
2453 of any of the test patterns, which is the same behaviour as GNU grep.
2455 7. A pattern that could match an empty string could cause pcregrep to loop; it
2456 doesn't make sense to accept an empty string match in pcregrep, so I have
2457 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2458 seems to be how GNU grep behaves. [But see later change 40 for release
2459 8.33.]
2461 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2462 start or after a newline", because the conditional assertion was not being
2463 correctly handled. The rule now is that both the assertion and what follows
2464 in the first alternative must satisfy the test.
2466 9. If auto-callout was enabled in a pattern with a conditional group whose
2467 condition was an assertion, PCRE could crash during matching, both with
2468 pcre_exec() and pcre_dfa_exec().
2470 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2471 used for matching.
2473 11. Unicode property support in character classes was not working for
2474 characters (bytes) greater than 127 when not in UTF-8 mode.
2476 12. Added the -M command line option to pcretest.
2478 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2480 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2482 16. Added comments and documentation about mis-use of no_arg in the C++
2483 wrapper.
2485 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2486 from Martin Jerabek that uses macro names for all relevant character and
2487 string constants.
2489 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2490 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2491 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2492 these, but not everybody uses configure.
2494 19. A conditional group that had only one branch was not being correctly
2495 recognized as an item that could match an empty string. This meant that an
2496 enclosing group might also not be so recognized, causing infinite looping
2497 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2498 with the subject "ab", where knowledge that the repeated group can match
2499 nothing is needed in order to break the loop.
2501 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2502 exec(), but without supplying a callout function, matching went wrong.
2504 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2505 leak if the size of the offset vector was greater than 30. When the vector
2506 is smaller, the saved offsets during recursion go onto a local stack
2507 vector, but for larger vectors malloc() is used. It was failing to free
2508 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2509 error, in fact).
2511 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2512 heapframe that is used only when UTF-8 support is enabled. This caused no
2513 problem, but was untidy.
2515 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2516 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2517 included within another project.
2519 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2520 slightly modified by me:
2522 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2523 not building pcregrep.
2525 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2526 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2528 25. Forward references, both numeric and by name, in patterns that made use of
2529 duplicate group numbers, could behave incorrectly or give incorrect errors,
2530 because when scanning forward to find the reference group, PCRE was not
2531 taking into account the duplicate group numbers. A pattern such as
2532 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2534 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2535 the feature of a custom pointer more persuasive (as requested by a user).
2537 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2538 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2539 pre-defined.
2541 28. Added support for (*UTF8) at the start of a pattern.
2543 29. Arrange for flags added by the "release type" setting in CMake to be shown
2544 in the configuration summary.
2547 Version 7.8 05-Sep-08
2548 ---------------------
2550 1. Replaced UCP searching code with optimized version as implemented for Ad
2551 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2552 stage table and inline lookup instead of a function, giving speed ups of 2
2553 to 5 times on some simple patterns that I tested. Permission was given to
2554 distribute the MultiStage2.py script that generates the tables (it's not in
2555 the tarball, but is in the Subversion repository).
2557 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2558 scripts.
2560 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2561 a group with a zero qualifier. The result of the study could be incorrect,
2562 or the function might crash, depending on the pattern.
2564 4. Caseless matching was not working for non-ASCII characters in back
2565 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2566 It now works when Unicode Property Support is available.
2568 5. In pcretest, an escape such as \x{de} in the data was always generating
2569 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2570 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2571 truncation.
2573 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2575 7. Added two (int) casts to pcregrep when printing the difference of two
2576 pointers, in case they are 64-bit values.
2578 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2579 test 2 if it fails.
2581 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2582 and a #define of that name to empty if it is not externally set. This is to
2583 allow users of MSVC to set it if necessary.
2585 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2586 the convenience functions in the pcre_get.c source file.
2588 11. An option change at the start of a pattern that had top-level alternatives
2589 could cause overwriting and/or a crash. This command provoked a crash in
2590 some environments:
2592 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2594 This potential security problem was recorded as CVE-2008-2371.
2596 12. For a pattern where the match had to start at the beginning or immediately
2597 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2598 pcre_dfa_exec() could read past the end of the passed subject if there was
2599 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2600 pcretest so that it places the subject at the end of its malloc-ed buffer.
2602 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2603 exec() might read past the end of the data buffer in UTF-8 mode.
2605 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2606 the data contained the byte 0x85 as part of a UTF-8 character within its
2607 first line. This applied both to normal and DFA matching.
2609 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2610 /^[^d]*?$/8 failed to match "abc".
2612 16. Added a missing copyright notice to pcrecpp_internal.h.
2614 17. Make it more clear in the documentation that values returned from
2615 pcre_exec() in ovector are byte offsets, not character counts.
2617 18. Tidied a few places to stop certain compilers from issuing warnings.
2619 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2620 supplied by Stefan Weber. I made a further small update for 7.8 because
2621 there is a change of source arrangements: the pcre_searchfuncs.c module is
2622 replaced by pcre_ucd.c.
2625 Version 7.7 07-May-08
2626 ---------------------
2628 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2629 a string to a long long, pretend we don't even have a long long." This is
2630 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2632 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2633 pre-7.6 versions, which defined a global no_arg variable instead of putting
2634 it in the RE class. (See also #8 below.)
2636 3. Remove a line of dead code, identified by coverity and reported by Nuno
2637 Lopes.
2639 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2641 (1) The include/exclude patterns were being applied to the whole pathnames
2642 of files, instead of just to the final components.
2644 (2) If there was more than one level of directory, the subdirectories were
2645 skipped unless they satisfied the include/exclude conditions. This is
2646 inconsistent with GNU grep (and could even be seen as contrary to the
2647 pcregrep specification - which I improved to make it absolutely clear).
2648 The action now is always to scan all levels of directory, and just
2649 apply the include/exclude patterns to regular files.
2651 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2652 --exclude_dir in the tests to avoid scanning .svn directories.
2654 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2655 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2656 doesn't support NULs in patterns.
2658 7. Added some missing "const"s to declarations of static tables in
2659 pcre_compile.c and pcre_dfa_exec.c.
2661 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2662 caused by fix #2 above. (Subsequently also a second patch to fix the
2663 first patch. And a third patch - this was a messy problem.)
2665 9. Applied Craig's patch to remove the use of push_back().
2667 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2668 matching function regexec().
2670 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2671 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2672 references. PCRE supports relative numbers with this syntax (I don't think
2673 Oniguruma does).
2675 12. Previously, a group with a zero repeat such as (...){0} was completely
2676 omitted from the compiled regex. However, this means that if the group
2677 was called as a subroutine from elsewhere in the pattern, things went wrong
2678 (an internal error was given). Such groups are now left in the compiled
2679 pattern, with a new opcode that causes them to be skipped at execution
2680 time.
2682 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2683 to the way PCRE behaves:
2685 (a) A lone ] character is dis-allowed (Perl treats it as data).
2687 (b) A back reference to an unmatched subpattern matches an empty string
2688 (Perl fails the current match path).
2690 (c) A data ] in a character class must be notated as \] because if the
2691 first data character in a class is ], it defines an empty class. (In
2692 Perl it is not possible to have an empty class.) The empty class []
2693 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2694 The negative empty class [^] matches any one character, independently
2695 of the DOTALL setting.
2697 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2698 non-existent subpattern following a character class starting with ']' and
2699 containing () gave an internal compiling error instead of "reference to
2700 non-existent subpattern". Fortunately, when the pattern did exist, the
2701 compiled code was correct. (When scanning forwards to check for the
2702 existence of the subpattern, it was treating the data ']' as terminating
2703 the class, so got the count wrong. When actually compiling, the reference
2704 was subsequently set up correctly.)
2706 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2707 it was being rejected as not supported by pcre_dfa_exec(), even though
2708 other assertions are supported. I have made pcre_dfa_exec() support
2709 (*FAIL).
2711 16. The implementation of 13c above involved the invention of a new opcode,
2712 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2713 cannot be changed at match time, I realized I could make a small
2714 improvement to matching performance by compiling OP_ALLANY instead of
2715 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2716 on the OP_ANY path.
2718 17. Compiling pcretest on Windows with readline support failed without the
2719 following two fixes: (1) Make the unistd.h include conditional on
2720 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2722 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2723 ncurses library to be included for pcretest when ReadLine support is
2724 requested, but also to allow for it to be overridden. This patch came from
2725 Daniel Bergström.
2727 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2728 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2729 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2730 spotting this.
2733 Version 7.6 28-Jan-08
2734 ---------------------
2736 1. A character class containing a very large number of characters with
2737 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2738 overflow.
2740 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2741 HAVE_LONG_LONG is not defined.
2743 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2744 bring them up to date and include new features. This patch includes:
2746 - Fixed PH's badly added libz and libbz2 support.
2747 - Fixed a problem with static linking.
2748 - Added pcredemo. [But later removed - see 7 below.]
2749 - Fixed dftables problem and added an option.
2750 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2752 - Added readline support for pcretest.
2753 - Added an listing of the option settings after cmake has run.
2755 4. A user submitted a patch to Makefile that makes it easy to create
2756 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2757 Makefile.am that cause it to include this special target, without
2758 affecting anything else. Note that the same mingw target plus all
2759 the other distribution libraries and programs are now supported
2760 when configuring with CMake (see 6 below) instead of with
2761 Configure/Make.
2763 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2764 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2765 exported in the Windows port". It has not yet been confirmed that the patch
2766 solves the problem, but it does no harm.
2768 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2769 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2770 with CMake, and also correct the comment about stack recursion.
2772 7. Remove the automatic building of pcredemo from the ./configure system and
2773 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2774 of a program that users should build themselves after PCRE is installed, so
2775 building it automatically is not really right. What is more, it gave
2776 trouble in some build environments.
2778 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2781 Version 7.5 10-Jan-08
2782 ---------------------
2784 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2785 values in parens when parsing an RE using the C++ wrapper."
2787 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2788 Characters greater than 255 were excluded from the class instead of being
2789 included.
2791 3. The same bug as (2) above applied to negated POSIX classes such as
2792 [:^space:].
2794 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2795 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2796 I have changed it.
2798 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2799 first named subpattern) and a construct such as (?&a) would reference the
2800 first named subpattern whose name started with "a" (in other words, the
2801 length check was missing). Both these problems are fixed. "Subpattern name
2802 expected" is now given for (?&) (a zero-length name), and this patch also
2803 makes it give the same error for \k'' (previously it complained that that
2804 was a reference to a non-existent subpattern).
2806 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2807 this is right because (?- can be followed by option settings as well as by
2808 digits. I have, however, made the messages clearer.
2810 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2811 than the number used in the conditional) now cause a compile-time error.
2812 This is actually not compatible with Perl, which accepts such patterns, but
2813 treats the conditional as always being FALSE (as PCRE used to), but it
2814 seems to me that giving a diagnostic is better.
2816 8. Change "alphameric" to the more common word "alphanumeric" in comments
2817 and messages.
2819 9. Fix two occurrences of "backslash" in comments that should have been
2820 "backspace".
2822 10. Remove two redundant lines of code that can never be obeyed (their function
2823 was moved elsewhere).
2825 11. The program that makes PCRE's Unicode character property table had a bug
2826 which caused it to generate incorrect table entries for sequences of
2827 characters that have the same character type, but are in different scripts.
2828 It amalgamated them into a single range, with the script of the first of
2829 them. In other words, some characters were in the wrong script. There were
2830 thirteen such cases, affecting characters in the following ranges:
2832 U+002b0 - U+002c1
2833 U+0060c - U+0060d
2834 U+0061e - U+00612
2835 U+0064b - U+0065e
2836 U+0074d - U+0076d
2837 U+01800 - U+01805
2838 U+01d00 - U+01d77
2839 U+01d9b - U+01dbf
2840 U+0200b - U+0200f
2841 U+030fc - U+030fe
2842 U+03260 - U+0327f
2843 U+0fb46 - U+0fbb1
2844 U+10450 - U+1049d
2846 12. The -o option (show only the matching part of a line) for pcregrep was not
2847 compatible with GNU grep in that, if there was more than one match in a
2848 line, it showed only the first of them. It now behaves in the same way as
2849 GNU grep.
2851 13. If the -o and -v options were combined for pcregrep, it printed a blank
2852 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2853 does the same. The return code can be used to tell if there were any
2854 non-matching lines.
2856 14. Added --file-offsets and --line-offsets to pcregrep.
2858 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2859 infinitely looping recursion. The bug was that positive lookaheads were not
2860 being skipped when checking for a possible empty match (negative lookaheads
2861 and both kinds of lookbehind were skipped).
2863 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2864 inclusion of <windows.h> to before rather than after the definition of
2865 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2867 17. Specifying a possessive quantifier with a specific limit for a Unicode
2868 character property caused pcre_compile() to compile bad code, which led at
2869 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2870 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2871 caused the error; without that there was no problem.
2873 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2875 19. Added --enable-pcretest-libreadline.
2877 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2878 RE::GlobalReplace(). As a result, the number of replacements returned was
2879 double what it should be. I removed one of the increments, but Craig sent a
2880 later patch that removed the other one (the right fix) and added unit tests
2881 that check the return values (which was not done before).
2883 21. Several CMake things:
2885 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2886 the names libpcre and libpcreposix, not just pcre and pcreposix.
2888 (2) The above change means that pcretest and pcregrep are now correctly
2889 linked with the newly-built libraries, not previously installed ones.
2893 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2894 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2895 UTF-8 newline character). The key issue is that the pattern starts .*;
2896 this means that the match must be either at the beginning, or after a
2897 newline. The bug was in the code for advancing after a failed match and
2898 checking that the new position followed a newline. It was not taking
2899 account of UTF-8 characters correctly.
2901 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2902 character classes. PCRE was not treating the sequence [:...:] as a
2903 character class unless the ... were all letters. Perl, however, seems to
2904 allow any characters between [: and :], though of course it rejects as
2905 unknown any "names" that contain non-letters, because all the known class
2906 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2907 for example, whereas PCRE did not - it did not recognize a POSIX character
2908 class. This seemed a bit dangerous, so the code has been changed to be
2909 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2910 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2911 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2912 Perl does, and where it didn't before.
2914 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2915 Windows environments %n is disabled by default.
2918 Version 7.4 21-Sep-07
2919 ---------------------
2921 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2922 means that a class such as [\s] counted as "explicit reference to CR or
2923 LF". That isn't really right - the whole point of the change was to try to
2924 help when there was an actual mention of one of the two characters. So now
2925 the change happens only if \r or \n (or a literal CR or LF) character is
2926 encountered.
2928 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2929 of both had grown to the point where there were only 3 bits left.
2930 Fortunately, there was spare space in the data structure, and so I have
2931 moved the internal flags into a new 16-bit field to free up more option
2932 bits.
2934 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2935 but did not set the internal JCHANGED flag - either of these is enough to
2936 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2937 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2938 start sets both bits.
2940 4. Added options (at build time, compile time, exec time) to change \R from
2941 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2943 5. doc/pcresyntax.html was missing from the distribution.
2945 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2946 compatibility, even though it is no longer used.
2948 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2949 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2950 windows.h file is present (where different names are used). [This was
2951 reversed later after testing - see 16 below.]
2953 8. Changed all #include <config.h> to #include "config.h". There were also
2954 some further <pcre.h> cases that I changed to "pcre.h".
2956 9. When pcregrep was used with the --colour option, it missed the line ending
2957 sequence off the lines that it output.
2959 10. It was pointed out to me that arrays of string pointers cause lots of
2960 relocations when a shared library is dynamically loaded. A technique of
2961 using a single long string with a table of offsets can drastically reduce
2962 these. I have refactored PCRE in four places to do this. The result is
2963 dramatic:
2965 Originally: 290
2966 After changing UCP table: 187
2967 After changing error message table: 43
2968 After changing table of "verbs" 36
2969 After changing table of Posix names 22
2971 Thanks to the folks working on Gregex for glib for this insight.
2973 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2974 unicode-properties was also set.
2976 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2978 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2979 checked only for CRLF.
2981 14. Added casts to pcretest.c to avoid compiler warnings.
2983 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2985 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2986 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2987 entirely. This removes changes made in 7 above.
2989 17. The CMake files have been updated, and there is now more information about
2990 building with CMake in the NON-UNIX-USE document.
2993 Version 7.3 28-Aug-07
2994 ---------------------
2996 1. In the rejigging of the build system that eventually resulted in 7.1, the
2997 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2998 brackets there is not right, since it causes compilers to look for an
2999 installed pcre.h, not the version that is in the source that is being
3000 compiled (which of course may be different). I have changed it back to:
3002 #include "pcre.h"
3004 I have a vague recollection that the change was concerned with compiling in
3005 different directories, but in the new build system, that is taken care of
3006 by the VPATH setting the Makefile.
3008 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3009 when the subject happened to end in the byte 0x85 (e.g. if the last
3010 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3011 characters but of course it shouldn't be taken as a newline when it is part
3012 of another character. The bug was that, for an unlimited repeat of . in
3013 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3014 characters when looking for a newline.
3016 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3018 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3019 in debug output.
3021 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3022 long printing in the pcrecpp unittest when running under MinGW.
3024 6. ESC_K was left out of the EBCDIC table.
3026 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3027 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3028 limit also applies to "virtual nesting" when a pattern is recursive, and in
3029 this case 1000 isn't so big. I have been able to remove this limit at the
3030 expense of backing off one optimization in certain circumstances. Normally,
3031 when pcre_exec() would call its internal match() function recursively and
3032 immediately return the result unconditionally, it uses a "tail recursion"
3033 feature to save stack. However, when a subpattern that can match an empty
3034 string has an unlimited repetition quantifier, it no longer makes this
3035 optimization. That gives it a stack frame in which to save the data for
3036 checking that an empty string has been matched. Previously this was taken
3037 from the 1000-entry workspace that had been reserved. So now there is no
3038 explicit limit, but more stack is used.
3040 8. Applied Daniel's patches to solve problems with the import/export magic
3041 syntax that is required for Windows, and which was going wrong for the
3042 pcreposix and pcrecpp parts of the library. These were overlooked when this
3043 problem was solved for the main library.
3045 9. There were some crude static tests to avoid integer overflow when computing
3046 the size of patterns that contain repeated groups with explicit upper
3047 limits. As the maximum quantifier is 65535, the maximum group length was
3048 set at 30,000 so that the product of these two numbers did not overflow a
3049 32-bit integer. However, it turns out that people want to use groups that
3050 are longer than 30,000 bytes (though not repeat them that many times).
3051 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3052 made it possible to implement the integer overflow checks in a much more
3053 dynamic way, which I have now done. The artificial limitation on group
3054 length has been removed - we now have only the limit on the total length of
3055 the compiled pattern, which depends on the LINK_SIZE setting.
3057 10. Fixed a bug in the documentation for get/copy named substring when
3058 duplicate names are permitted. If none of the named substrings are set, the
3059 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3060 empty string.
3062 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3063 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3064 because the ] is interpreted as the first data character and the
3065 terminating ] is not found. PCRE has been made compatible with Perl in this
3066 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3067 cause memory overwriting.
3069 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3070 string has been matched (to stop an infinite loop). It was not recognizing
3071 a conditional subpattern that could match an empty string if that
3072 subpattern was within another subpattern. For example, it looped when
3073 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3074 condition was not nested. This bug has been fixed.
3076 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3077 past the start of the subject in the presence of bytes with the top bit
3078 set, for example "\x8aBCD".
3080 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3081 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3083 14. Optimized (?!) to (*FAIL).
3085 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3086 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3087 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3088 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3089 does: it's just the validity check that is more restrictive.
3091 16. Inserted checks for integer overflows during escape sequence (backslash)
3092 processing, and also fixed erroneous offset values for syntax errors during
3093 backslash processing.
3095 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3096 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3098 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3099 caused an overrun.
3101 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3102 something other than just ASCII characters) inside a group that had an
3103 unlimited repeat caused a loop at compile time (while checking to see
3104 whether the group could match an empty string).
3106 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3107 [\P{Any}] did so. (Error in the code for printing property names.)
3109 21. An orphan \E inside a character class could cause a crash.
3111 22. A repeated capturing bracket such as (A)? could cause a wild memory
3112 reference during compilation.
3114 23. There are several functions in pcre_compile() that scan along a compiled
3115 expression for various reasons (e.g. to see if it's fixed length for look
3116 behind). There were bugs in these functions when a repeated \p or \P was
3117 present in the pattern. These operators have additional parameters compared
3118 with \d, etc, and these were not being taken into account when moving along
3119 the compiled data. Specifically:
3121 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3122 length.
3124 (b) An item such as \pL+ within a repeated group could cause crashes or
3125 loops.
3127 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3128 "reference to non-existent subpattern" error.
3130 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3132 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3133 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3135 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3137 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3138 character were causing crashes (broken optimization).
3140 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3141 \p or \P) caused a compile-time loop.
3143 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3144 break. For example, the unstudied pattern [\r\n]A does not match the string
3145 "\r\nA" because change 7.0/46 below moves the current point on by two
3146 characters after failing to match at the start. However, the pattern \nA
3147 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3148 the same is true. There doesn't seem any very clean way out of this, but
3149 what I have chosen to do makes the common cases work: PCRE now takes note
3150 of whether there can be an explicit match for \r or \n anywhere in the
3151 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3152 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3153 pattern has explicit CR or LF references.
3155 29. Added (*CR) etc for changing newline setting at start of pattern.
3158 Version 7.2 19-Jun-07
3159 ---------------------
3161 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3162 which is apparently normally available under Windows.
3164 2. Re-jig the pcregrep tests with different newline settings in an attempt
3165 to make them independent of the local environment's newline setting.
3167 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3169 4. Some of the "internals" tests were previously cut out when the link size
3170 was not 2, because the output contained actual offsets. The recent new
3171 "Z" feature of pcretest means that these can be cut out, making the tests
3172 usable with all link sizes.
3174 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3175 stack recursion. This gives a massive performance boost under BSD, but just
3176 a small improvement under Linux. However, it saves one field in the frame
3177 in all cases.
3179 6. Added more features from the forthcoming Perl 5.10:
3181 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3182 recursion call. It refers to the nth most recently opened parentheses.
3184 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3185 to be opened parentheses.
3187 (c) Conditions that refer to capturing parentheses can be specified
3188 relatively, for example, (?(-2)... or (?(+3)...
3190 (d) \K resets the start of the current match so that everything before
3191 is not part of it.
3193 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3195 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3196 reference syntax.
3198 (g) (?| introduces a group in which the numbering of parentheses in each
3199 alternative starts with the same number.
3201 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3203 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3206 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3207 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3208 for detecting groups that can match an empty string.
3210 9. A pattern with a very large number of alternatives (more than several
3211 hundred) was running out of internal workspace during the pre-compile
3212 phase, where pcre_compile() figures out how much memory will be needed. A
3213 bit of new cunning has reduced the workspace needed for groups with
3214 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3215 workspace instead of running out of the 4096 that are available.
3217 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3219 11. Applied patch from Google to remove an optimization that didn't quite work.
3220 The report of the bug said:
3222 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3223 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3224 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3226 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3227 it matched the wrong number of bytes.
3230 Version 7.1 24-Apr-07
3231 ---------------------
3233 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3234 that is more "standard", making use of automake and other Autotools. There
3235 is some re-arrangement of the files and adjustment of comments consequent
3236 on this.
3238 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3239 for recursive directory scanning broke on some systems because the files
3240 are not scanned in any specific order and on different systems the order
3241 was different. A call to "sort" has been inserted into RunGrepTest for the
3242 approprate test as a short-term fix. In the longer term there may be an
3243 alternative.
3245 3. I had an email from Eric Raymond about problems translating some of PCRE's
3246 man pages to HTML (despite the fact that I distribute HTML pages, some
3247 people do their own conversions for various reasons). The problems
3248 concerned the use of low-level troff macros .br and .in. I have therefore
3249 removed all such uses from the man pages (some were redundant, some could
3250 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3251 HTML has been updated to handle .nf/.fi and to complain if it encounters
3252 .br or .in.
3254 4. Updated comments in configure.ac that get placed in config.h.in and also
3255 arranged for config.h to be included in the distribution, with the name
3256 config.h.generic, for the benefit of those who have to compile without
3257 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3259 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3260 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3261 makevp.bat for the current PCRE, using the additional files
3262 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3264 6. A Windows user reported a minor discrepancy with test 2, which turned out
3265 to be caused by a trailing space on an input line that had got lost in his
3266 copy. The trailing space was an accident, so I've just removed it.
3268 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3269 that is needed.
3271 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3272 as "const" (a) because they are and (b) because it helps the PHP
3273 maintainers who have recently made a script to detect big data structures
3274 in the php code that should be moved to the .rodata section. I remembered
3275 to update Builducptable as well, so it won't revert if ucptable.h is ever
3276 re-created.
3278 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3279 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3280 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3281 support is not required. This saves 1.5-2K of code, which is important in
3282 some applications.
3284 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3285 so as not to refer to the tables, even though these functions will never be
3286 called when UTF-8 support is disabled. Otherwise there are problems with a
3287 shared library.
3289 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3291 (a) It was defining its arguments as char * instead of void *.
3293 (b) It was assuming that all moves were upwards in memory; this was true
3294 a long time ago when I wrote it, but is no longer the case.
3296 The emulated memove() is provided for those environments that have neither
3297 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3298 is clearly not the case, as these two bugs were recently reported.
3300 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3301 and Detrail to create the HTML documentation, the .txt form of the man
3302 pages, and it removes trailing spaces from listed files. It also creates
3303 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3304 case, it wraps all the #defines with #ifndefs. This script should be run
3305 before "make dist".
3307 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3308 with Unicode property support.
3310 (a) For a maximizing quantifier, if the two different cases of the
3311 character were of different lengths in their UTF-8 codings (there are
3312 some cases like this - I found 11), and the matching function had to
3313 back up over a mixture of the two cases, it incorrectly assumed they
3314 were both the same length.
3316 (b) When PCRE was configured to use the heap rather than the stack for
3317 recursion during matching, it was not correctly preserving the data for
3318 the other case of a UTF-8 character when checking ahead for a match
3319 while processing a minimizing repeat. If the check also involved
3320 matching a wide character, but failed, corruption could cause an
3321 erroneous result when trying to check for a repeat of the original
3322 character.
3324 13. Some tidying changes to the testing mechanism:
3326 (a) The RunTest script now detects the internal link size and whether there
3327 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3328 values substituted by "configure". (The RunGrepTest script already did
3329 this for UTF-8.) The configure.ac script no longer substitutes the
3330 relevant variables.
3332 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3333 with length and offset values. This means that the output is different
3334 for different internal link sizes. Test 2 is skipped for link sizes
3335 other than 2 because of this, bypassing the problem. Unfortunately,
3336 there was also a test in test 3 (the locale tests) that used /B and
3337 failed for link sizes other than 2. Rather than cut the whole test out,
3338 I have added a new /Z option to pcretest that replaces the length and
3339 offset values with spaces. This is now used to make test 3 independent
3340 of link size. (Test 2 will be tidied up later.)
3342 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3343 segmentation fault instead of returning the appropriate error message.
3345 15. In multiline mode when the newline sequence was set to "any", the pattern
3346 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3347 This doesn't seem right; it now treats the CRLF combination as the line
3348 ending, and so does not match in that case. It's only a pattern such as ^$
3349 that would hit this one: something like ^ABC$ would have failed after \r
3350 and then tried again after \r\n.
3352 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3353 in an attempt to make files that differ only in their line terminators
3354 compare equal. This works on Linux.
3356 17. Under certain error circumstances pcregrep might try to free random memory
3357 as it exited. This is now fixed, thanks to valgrind.
3359 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3360 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3361 was because its rules for how to advance for /g after matching an empty
3362 string at the end of a line did not allow for this case. They now check for
3363 it specially.
3365 20. pcretest is supposed to handle patterns and data of any length, by
3366 extending its buffers when necessary. It was getting this wrong when the
3367 buffer for a data line had to be extended.
3369 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3370 CRLF as a newline sequence.
3372 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3373 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3374 I have nevertheless tidied it up.
3376 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3378 24. Added a man page for pcre-config.
3381 Version 7.0 19-Dec-06
3382 ---------------------
3384 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3385 moving to gcc 4.1.1.
3387 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3388 sys/time.h, which is documented as needed for this function. It doesn't
3389 seem to matter on Linux, but it showed up on some releases of OS X.
3391 3. It seems that there are systems where bytes whose values are greater than
3392 127 match isprint() in the "C" locale. The "C" locale should be the
3393 default when a C program starts up. In most systems, only ASCII printing
3394 characters match isprint(). This difference caused the output from pcretest
3395 to vary, making some of the tests fail. I have changed pcretest so that:
3397 (a) When it is outputting text in the compiled version of a pattern, bytes
3398 other than 32-126 are always shown as hex escapes.
3400 (b) When it is outputting text that is a matched part of a subject string,
3401 it does the same, unless a different locale has been set for the match
3402 (using the /L modifier). In this case, it uses isprint() to decide.
3404 4. Fixed a major bug that caused incorrect computation of the amount of memory
3405 required for a compiled pattern when options that changed within the
3406 pattern affected the logic of the preliminary scan that determines the
3407 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3408 that the computed length was too small. The symptoms of this bug were
3409 either the PCRE error "internal error: code overflow" from pcre_compile(),
3410 or a glibc crash with a message such as "pcretest: free(): invalid next
3411 size (fast)". Examples of patterns that provoked this bug (shown in
3412 pcretest format) are:
3414 /(?-x: )/x
3415 /(?x)(?-x: \s*#\s*)/
3416 /((?i)[\x{c0}])/8
3417 /(?i:[\x{c0}])/8
3419 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3420 is now done differently.
3422 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3423 wrapper classes; (b) implement a new function in the C++ scanner that is
3424 more efficient than the old way of doing things because it avoids levels of
3425 recursion in the regex matching; (c) add a paragraph to the documentation
3426 for the FullMatch() function.
3428 6. The escape sequence \n was being treated as whatever was defined as
3429 "newline". Not only was this contrary to the documentation, which states
3430 that \n is character 10 (hex 0A), but it also went horribly wrong when
3431 "newline" was defined as CRLF. This has been fixed.
3433 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3434 was being set to -1 for the "end of line" case (supposedly a value that no
3435 character can have). Though this value is never used (the check for end of
3436 line is "zero bytes in current character"), it caused compiler complaints.
3437 I've changed it to 0xffffffff.
3439 8. In pcre_version.c, the version string was being built by a sequence of
3440 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3441 string (as it is for production releases) called a macro with an empty
3442 argument. The C standard says the result of this is undefined. The gcc
3443 compiler treats it as an empty string (which was what was wanted) but it is
3444 reported that Visual C gives an error. The source has been hacked around to
3445 avoid this problem.
3447 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3448 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3449 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3450 of them did).
3452 10. Originally, pcretest opened its input and output without "b"; then I was
3453 told that "b" was needed in some environments, so it was added for release
3454 5.0 to both the input and output. (It makes no difference on Unix-like
3455 systems.) Later I was told that it is wrong for the input on Windows. I've
3456 now abstracted the modes into two macros, to make it easier to fiddle with
3457 them, and removed "b" from the input mode under Windows.
3459 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3461 12. Added -help and --help to pcretest as an official way of being reminded
3462 of the options.
3464 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3465 and pcrecpp.cc because they annoy compilers at high warning levels.
3467 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3469 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3470 scripts are not C programs :-) and which was not noticed because it works
3471 on Linux.
3473 16. pcretest is supposed to handle any length of pattern and data line (as one
3474 line or as a continued sequence of lines) by extending its input buffer if
3475 necessary. This feature was broken for very long pattern lines, leading to
3476 a string of junk being passed to pcre_compile() if the pattern was longer
3477 than about 50K.
3479 17. I have done a major re-factoring of the way pcre_compile() computes the
3480 amount of memory needed for a compiled pattern. Previously, there was code
3481 that made a preliminary scan of the pattern in order to do this. That was
3482 OK when PCRE was new, but as the facilities have expanded, it has become
3483 harder and harder to keep it in step with the real compile phase, and there
3484 have been a number of bugs (see for example, 4 above). I have now found a
3485 cunning way of running the real compile function in a "fake" mode that
3486 enables it to compute how much memory it would need, while actually only
3487 ever using a few hundred bytes of working memory and without too many
3488 tests of the mode. This should make future maintenance and development
3489 easier. A side effect of this work is that the limit of 200 on the nesting
3490 depth of parentheses has been removed (though this was never a serious
3491 limitation, I suspect). However, there is a downside: pcre_compile() now
3492 runs more slowly than before (30% or more, depending on the pattern). I
3493 hope this isn't a big issue. There is no effect on runtime performance.
3495 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3496 newline (only possible for the last line of a file) and it was a
3497 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3499 19. Added additional timing features to pcretest. (1) The -tm option now times
3500 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3501 separate command line item, by a number that specifies the number of
3502 repeats to use when timing. The default is 50000; this gives better
3503 precision, but takes uncomfortably long for very large patterns.
3505 20. Extended pcre_study() to be more clever in cases where a branch of a
3506 subpattern has no definite first character. For example, (a*|b*)[cd] would
3507 previously give no result from pcre_study(). Now it recognizes that the
3508 first character must be a, b, c, or d.
3510 21. There was an incorrect error "recursive call could loop indefinitely" if
3511 a subpattern (or the entire pattern) that was being tested for matching an
3512 empty string contained only one non-empty item after a nested subpattern.
3513 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3514 incorrectly, because the \d was being skipped in the check.
3516 22. The pcretest program now has a new pattern option /B and a command line
3517 option -b, which is equivalent to adding /B to every pattern. This causes
3518 it to show the compiled bytecode, without the additional information that
3519 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3520 is the same as /B/I).
3522 23. A new optimization is now able automatically to treat some sequences such
3523 as a*b as a*+b. More specifically, if something simple (such as a character
3524 or a simple class like \d) has an unlimited quantifier, and is followed by
3525 something that cannot possibly match the quantified thing, the quantifier
3526 is automatically "possessified".
3528 24. A recursive reference to a subpattern whose number was greater than 39
3529 went wrong under certain circumstances in UTF-8 mode. This bug could also
3530 have affected the operation of pcre_study().
3532 25. Realized that a little bit of performance could be had by replacing
3533 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3535 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3537 27. Possessive quantifiers such as a++ were previously implemented by turning
3538 them into atomic groups such as ($>a+). Now they have their own opcodes,
3539 which improves performance. This includes the automatically created ones
3540 from 23 above.
3542 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3543 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3544 the first matched character to be a colon. This applied both to named and
3545 numbered groups.
3547 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3549 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3550 building PCRE on Windows easier, so I have included it in the distribution.
3552 31. There is now a check in pcretest against a ridiculously large number being
3553 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3554 loop, the loop is abandoned.
3556 32. Forward references to subpatterns in conditions such as (?(2)...) where
3557 subpattern 2 is defined later cause pcre_compile() to search forwards in
3558 the pattern for the relevant set of parentheses. This search went wrong
3559 when there were unescaped parentheses in a character class, parentheses
3560 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3562 33. "Subroutine" calls and backreferences were previously restricted to
3563 referencing subpatterns earlier in the regex. This restriction has now
3564 been removed.
3566 34. Added a number of extra features that are going to be in Perl 5.10. On the
3567 whole, these are just syntactic alternatives for features that PCRE had
3568 previously implemented using the Python syntax or my own invention. The
3569 other formats are all retained for compatibility.
3571 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3572 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3573 also .NET compatible.
3575 (b) A recursion or subroutine call to a named group can now be defined as
3576 (?&name) as well as (?P>name).
3578 (c) A backreference to a named group can now be defined as \k<name> or
3579 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3580 5.10, are also .NET compatible.
3582 (d) A conditional reference to a named group can now use the syntax
3583 (?(<name>) or (?('name') as well as (?(name).
3585 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3586 groups (named and numbered) that are never evaluated inline, but can be
3587 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3588 is always false. There may be only one alternative in such a group.
3590 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3591 as the simple (?(R). The condition is true only if the most recent
3592 recursion is that of the given number or name. It does not search out
3593 through the entire recursion stack.
3595 (g) The escape \gN or \g{N} has been added, where N is a positive or
3596 negative number, specifying an absolute or relative reference.
3598 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3599 some "unreachable code" warnings.
3601 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3602 things, this adds five new scripts.
3604 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3605 There were also incompatibilities regarding the handling of \Q..\E inside
3606 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3607 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3609 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3610 matches an empty string, and forcibly breaks the loop. There were bugs in
3611 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3612 against aaaa the result was just "a" rather than "aaaa", for example. Two
3613 separate and independent bugs (that affected different cases) have been
3614 fixed.
3616 39. Refactored the code to abolish the use of different opcodes for small
3617 capturing bracket numbers. This is a tidy that I avoided doing when I
3618 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3619 The new approach is not only tidier, it makes it possible to reduce the
3620 memory needed to fix the previous bug (38).
3622 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3623 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3624 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3625 mode.
3627 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3628 report.
3630 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3631 copy construction and assignment in the C++ wrapper.
3633 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3634 couple of bugs that could have given wrong results in the "--newline=crlf"
3635 case.
3637 44. Added a number of casts and did some reorganization of signed/unsigned int
3638 variables following suggestions from Dair Grant. Also renamed the variable
3639 "this" as "item" because it is a C++ keyword.
3641 45. Arranged for dftables to add
3643 #include "pcre_internal.h"
3645 to pcre_chartables.c because without it, gcc 4.x may remove the array
3646 definition from the final binary if PCRE is built into a static library and
3647 dead code stripping is activated.
3649 46. For an unanchored pattern, if a match attempt fails at the start of a
3650 newline sequence, and the newline setting is CRLF or ANY, and the next two
3651 characters are CRLF, advance by two characters instead of one.
3654 Version 6.7 04-Jul-06
3655 ---------------------
3657 1. In order to handle tests when input lines are enormously long, pcretest has
3658 been re-factored so that it automatically extends its buffers when
3659 necessary. The code is crude, but this _is_ just a test program. The
3660 default size has been increased from 32K to 50K.
3662 2. The code in pcre_study() was using the value of the re argument before
3663 testing it for NULL. (Of course, in any sensible call of the function, it
3664 won't be NULL.)
3666 3. The memmove() emulation function in pcre_internal.h, which is used on
3667 systems that lack both memmove() and bcopy() - that is, hardly ever -
3668 was missing a "static" storage class specifier.
3670 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3671 containing an extended class (one that cannot be represented by a bitmap
3672 because it contains high-valued characters or Unicode property items, e.g.
3673 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3674 pattern, but PCRE should not loop if you do not (it no longer does).
3675 [Detail: two cases were found: (a) a repeated subpattern containing an
3676 extended class; (b) a recursive reference to a subpattern that followed a
3677 previous extended class. It wasn't skipping over the extended class
3678 correctly when UTF-8 mode was not set.]
3680 5. A negated single-character class was not being recognized as fixed-length
3681 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3682 compile error "lookbehind assertion is not fixed length".
3684 6. The RunPerlTest auxiliary script was showing an unexpected difference
3685 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3686 write a Perl script that can interpret lines of an input file either as
3687 byte characters or as UTF-8, which is what "perltest" was being required to
3688 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3689 can't do is switch easily at run time between having the "use utf8;" pragma
3690 or not. In the end, I fudged it by using the RunPerlTest script to insert
3691 "use utf8;" explicitly for the UTF-8 tests.
3693 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3694 the end of the subject string, contrary to the documentation and to what
3695 Perl does. This was true of both matching functions. Now it matches only at
3696 the start of the subject and immediately after *internal* newlines.
3698 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3699 a pointer to an int instead of a pointer to an unsigned long int. This
3700 caused problems on 64-bit systems.
3702 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3703 instance of the 'standard' template library not being so standard".
3705 10. There was no check on the number of named subpatterns nor the maximum
3706 length of a subpattern name. The product of these values is used to compute
3707 the size of the memory block for a compiled pattern. By supplying a very
3708 long subpattern name and a large number of named subpatterns, the size
3709 computation could be caused to overflow. This is now prevented by limiting
3710 the length of names to 32 characters, and the number of named subpatterns
3711 to 10,000.
3713 11. Subpatterns that are repeated with specific counts have to be replicated in
3714 the compiled pattern. The size of memory for this was computed from the
3715 length of the subpattern and the repeat count. The latter is limited to
3716 65535, but there was no limit on the former, meaning that integer overflow
3717 could in principle occur. The compiled length of a repeated subpattern is
3718 now limited to 30,000 bytes in order to prevent this.
3720 12. Added the optional facility to have named substrings with the same name.
3722 13. Added the ability to use a named substring as a condition, using the
3723 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3724 are numbers (not recommended). Forward references are permitted.
3726 14. Added forward references in named backreferences (if you see what I mean).
3728 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3729 pattern could run off the end of the subject. For example, the pattern
3730 "(?s)(.{1,5})"8 did this with the subject "ab".
3732 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3733 PCRE_CASELESS was set when matching characters that were quantified with ?
3734 or *.
3736 17. A character class other than a single negated character that had a minimum
3737 but no maximum quantifier - for example [ab]{6,} - was not handled
3738 correctly by pce_dfa_exec(). It would match only one character.
3740 18. A valid (though odd) pattern that looked like a POSIX character
3741 class but used an invalid character after [ (for example [[,abc,]]) caused
3742 pcre_compile() to give the error "Failed: internal error: code overflow" or
3743 in some cases to crash with a glibc free() error. This could even happen if
3744 the pattern terminated after [[ but there just happened to be a sequence of
3745 letters, a binary zero, and a closing ] in the memory that followed.
3747 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3748 over the years. Originally (before any Unicode support), just the bottom 8
3749 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3750 output from "man perlunicode" includes this:
3752 The regular expression compiler produces polymorphic opcodes. That
3753 is, the pattern adapts to the data and automatically switches to
3754 the Unicode character scheme when presented with Unicode data--or
3755 instead uses a traditional byte scheme when presented with byte
3756 data.
3758 Sadly, a wide octal escape does not cause a switch, and in a string with
3759 no other multibyte characters, these octal escapes are treated as before.
3760 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3761 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3762 Unicode string.
3764 I have not perpetrated such confusion in PCRE. Up till now, it took just
3765 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3766 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3767 translate to the appropriate multibyte character.
3769 29. Applied some refactoring to reduce the number of warnings from Microsoft
3770 and Borland compilers. This has included removing the fudge introduced
3771 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3772 a warning about an unused variable.
3774 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3775 characters since release 4.0, because Perl (from release 5.004) does not.
3776 [Or at least, is documented not to: some releases seem to be in conflict
3777 with the documentation.] However, when a pattern was studied with
3778 pcre_study() and all its branches started with \s, PCRE still included VT
3779 as a possible starting character. Of course, this did no harm; it just
3780 caused an unnecessary match attempt.
3782 22. Removed a now-redundant internal flag bit that recorded the fact that case
3783 dependency changed within the pattern. This was once needed for "required
3784 byte" processing, but is no longer used. This recovers a now-scarce options
3785 bit. Also moved the least significant internal flag bit to the most-
3786 significant bit of the word, which was not previously used (hangover from
3787 the days when it was an int rather than a uint) to free up another bit for
3788 the future.
3790 23. Added support for CRLF line endings as well as CR and LF. As well as the
3791 default being selectable at build time, it can now be changed at runtime
3792 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3793 specify that it is scanning data with non-default line endings.
3795 24. Changed the definition of CXXLINK to make it agree with the definition of
3796 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3798 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3799 recursions. This makes a big different to stack usage for some patterns.
3801 26. If a subpattern containing a named recursion or subroutine reference such
3802 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3803 the space required for the compiled pattern went wrong and gave too small a
3804 value. Depending on the environment, this could lead to "Failed: internal
3805 error: code overflow at offset 49" or "glibc detected double free or
3806 corruption" errors.
3808 27. Applied patches from Google (a) to support the new newline modes and (b) to
3809 advance over multibyte UTF-8 characters in GlobalReplace.
3811 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3812 difference for some implementation of PCRE in some Windows version.
3814 29. Added some extra testing facilities to pcretest:
3816 \q<number> in a data line sets the "match limit" value
3817 \Q<number> in a data line sets the "match recursion limt" value
3818 -S <number> sets the stack size, where <number> is in megabytes
3820 The -S option isn't available for Windows.
3823 Version 6.6 06-Feb-06
3824 ---------------------
3826 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3827 in pcreposix.h. I have copied the definition from pcre.h.
3829 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3830 because pcre.h is no longer a built file.
3832 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3833 not normally included in the compiled code.
3836 Version 6.5 01-Feb-06
3837 ---------------------
3839 1. When using the partial match feature with pcre_dfa_exec(), it was not
3840 anchoring the second and subsequent partial matches at the new starting
3841 point. This could lead to incorrect results. For example, with the pattern
3842 /1234/, partially matching against "123" and then "a4" gave a match.
3844 2. Changes to pcregrep:
3846 (a) All non-match returns from pcre_exec() were being treated as failures
3847 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3848 error message is output. Some extra information is given for the
3850 probably the only errors that are likely to be caused by users (by
3851 specifying a regex that has nested indefinite repeats, for instance).
3852 If there are more than 20 of these errors, pcregrep is abandoned.
3854 (b) A binary zero was treated as data while matching, but terminated the
3855 output line if it was written out. This has been fixed: binary zeroes
3856 are now no different to any other data bytes.
3858 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3859 used to set a locale for matching. The --locale=xxxx long option has
3860 been added (no short equivalent) to specify a locale explicitly on the
3861 pcregrep command, overriding the environment variables.
3863 (d) When -B was used with -n, some line numbers in the output were one less
3864 than they should have been.
3866 (e) Added the -o (--only-matching) option.
3868 (f) If -A or -C was used with -c (count only), some lines of context were
3869 accidentally printed for the final match.
3871 (g) Added the -H (--with-filename) option.
3873 (h) The combination of options -rh failed to suppress file names for files
3874 that were found from directory arguments.
3876 (i) Added the -D (--devices) and -d (--directories) options.
3878 (j) Added the -F (--fixed-strings) option.
3880 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3882 (l) Added the --colo(u)r option.
3884 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3885 is not present by default.
3887 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3888 items such as (?R) or (?1), when the recursion could match a number of
3889 alternatives. If it matched one of the alternatives, but subsequently,
3890 outside the recursion, there was a failure, the code tried to back up into
3891 the recursion. However, because of the way PCRE is implemented, this is not
3892 possible, and the result was an incorrect result from the match.
3894 In order to prevent this happening, the specification of recursion has
3895 been changed so that all such subpatterns are automatically treated as
3896 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3898 4. I had overlooked the fact that, in some locales, there are characters for
3899 which isalpha() is true but neither isupper() nor islower() are true. In
3900 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3901 and ordfeminine) are like this. This affected the treatment of \w and \W
3902 when they appeared in character classes, but not when they appeared outside
3903 a character class. The bit map for "word" characters is now created
3904 separately from the results of isalnum() instead of just taking it from the
3905 upper, lower, and digit maps. (Plus the underscore character, of course.)
3907 5. The above bug also affected the handling of POSIX character classes such as
3908 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3909 permanent tables. Instead, the bit maps for such a class were previously
3910 created as the appropriate unions of the upper, lower, and digit bitmaps.
3911 Now they are created by subtraction from the [[:word:]] class, which has
3912 its own bitmap.
3914 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3915 It is created by subtracting the vertical space characters (\x09, \x0a,
3916 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3917 subtraction was done in the overall bitmap for a character class, meaning
3918 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3919 be recognized. This bug has been fixed.
3921 7. Patches from the folks at Google:
3923 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3924 real life, but is still worth protecting against".
3926 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3927 regular expressions".
3929 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3930 have it.
3932 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3933 "configure" and the latter not, in order to fix a problem somebody had
3934 with compiling the Arg class on HP-UX.
3936 (e) Improve the error-handling of the C++ wrapper a little bit.
3938 (f) New tests for checking recursion limiting.
3940 8. The pcre_memmove() function, which is used only if the environment does not
3941 have a standard memmove() function (and is therefore rarely compiled),
3942 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3943 returning a result (though PCRE never actually uses the result).
3945 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3946 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3947 returned instead of calling malloc() with an overflowing number that would
3948 most likely cause subsequent chaos.
3950 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3952 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3953 with this option is matched, the nmatch and pmatch options of regexec() are
3954 ignored.
3956 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3957 provided in case anyone wants to the the POSIX interface with UTF-8
3958 strings.
3960 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3961 C++ linking (needed for some HP-UX environments).
3963 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3964 (unused parameter) and in the pcre_printint() function (omitted "default"
3965 switch label when the default is to do nothing).
3967 15. Added some code to make it possible, when PCRE is compiled as a C++
3968 library, to replace subject pointers for pcre_exec() with a smart pointer
3969 class, thus making it possible to process discontinuous strings.
3971 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3972 much the same function. They were added by different people who were trying
3973 to make PCRE easy to compile on non-Unix systems. It has been suggested
3974 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3975 for compiling on Windows systems. I have therefore replaced it with
3976 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3977 defaults to "extern" for C or "extern C" for C++, which works fine on
3978 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3979 SCOPE with something explicit in config.h. In addition:
3981 (a) pcreposix.h still had just "extern" instead of either of these macros;
3982 I have replaced it with PCRE_DATA_SCOPE.
3984 (b) Functions such as _pcre_xclass(), which are internal to the library,
3985 but external in the C sense, all had PCRE_EXPORT in their definitions.
3986 This is apparently wrong for the Windows case, so I have removed it.
3987 (It makes no difference on Unix-like systems.)
3989 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3990 of recursive calls to match(). This is different to MATCH_LIMIT because
3991 that limits the total number of calls to match(), not all of which increase
3992 the depth of recursion. Limiting the recursion depth limits the amount of
3993 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3994 when PCRE is compiled, and changed at run time. A patch from Google adds
3995 this functionality to the C++ interface.
3997 18. Changes to the handling of Unicode character properties:
3999 (a) Updated the table to Unicode 4.1.0.
4001 (b) Recognize characters that are not in the table as "Cn" (undefined).
4003 (c) I revised the way the table is implemented to a much improved format
4004 which includes recognition of ranges. It now supports the ranges that
4005 are defined in UnicodeData.txt, and it also amalgamates other
4006 characters into ranges. This has reduced the number of entries in the
4007 table from around 16,000 to around 3,000, thus reducing its size
4008 considerably. I realized I did not need to use a tree structure after
4009 all - a binary chop search is just as efficient. Having reduced the
4010 number of entries, I extended their size from 6 bytes to 8 bytes to
4011 allow for more data.
4013 (d) Added support for Unicode script names via properties such as \p{Han}.
4015 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4016 matching that character.
4018 20. When matching a repeated Unicode property with a minimum greater than zero,
4019 (for example \pL{2,}), PCRE could look past the end of the subject if it
4020 reached it while seeking the minimum number of characters. This could
4021 happen only if some of the characters were more than one byte long, because
4022 there is a check for at least the minimum number of bytes.
4024 21. Refactored the implementation of \p and \P so as to be more general, to
4025 allow for more different types of property in future. This has changed the
4026 compiled form incompatibly. Anybody with saved compiled patterns that use
4027 \p or \P will have to recompile them.
4029 22. Added "Any" and "L&" to the supported property types.
4031 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4032 but give a compile time error if the value is greater than 0xff.
4034 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4035 accidentally not being installed or uninstalled.
4037 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4038 made were to insert the current release number. This seemed silly, because
4039 it made things harder for people building PCRE on systems that don't run
4040 "configure". I have turned pcre.h into a distributed file, no longer built
4041 by "configure", with the version identification directly included. There is
4042 no longer a pcre.h.in file.
4044 However, this change necessitated a change to the pcre-config script as
4045 well. It is built from pcre-config.in, and one of the substitutions was the
4046 release number. I have updated configure.ac so that ./configure now finds
4047 the release number by grepping pcre.h.
4049 26. Added the ability to run the tests under valgrind.
4052 Version 6.4 05-Sep-05
4053 ---------------------
4055 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4056 "--" to be printed when multiple files were scanned, even when none of the
4057 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4058 consider it to be a bug, and have restored the previous behaviour.
4060 2. A couple of code tidies to get rid of compiler warnings.
4062 3. The pcretest program used to cheat by referring to symbols in the library
4063 whose names begin with _pcre_. These are internal symbols that are not
4064 really supposed to be visible externally, and in some environments it is
4065 possible to suppress them. The cheating is now confined to including
4066 certain files from the library's source, which is a bit cleaner.
4068 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4069 file's purpose clearer.
4071 5. Reorganized pcre_ucp_findchar().
4074 Version 6.3 15-Aug-05
4075 ---------------------
4077 1. The file libpcre.pc.in did not have general read permission in the tarball.
4079 2. There were some problems when building without C++ support:
4081 (a) If C++ support was not built, "make install" and "make test" still
4082 tried to test it.
4084 (b) There were problems when the value of CXX was explicitly set. Some
4085 changes have been made to try to fix these, and ...
4087 (c) --disable-cpp can now be used to explicitly disable C++ support.
4089 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4090 backslash in a target when C++ was disabled. This confuses some
4091 versions of "make", apparently. Using an intermediate variable solves
4092 this. (Same for CPP_LOBJ.)
4094 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4095 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4096 necessary on certain architectures.
4098 4. Added a setting of -export-symbols-regex to the link command to remove
4099 those symbols that are exported in the C sense, but actually are local
4100 within the library, and not documented. Their names all begin with
4101 "_pcre_". This is not a perfect job, because (a) we have to except some
4102 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4103 available (and never for static libraries). I have made a note to try to
4104 find a way round (a) in the future.
4107 Version 6.2 01-Aug-05
4108 ---------------------
4110 1. There was no test for integer overflow of quantifier values. A construction
4111 such as {1111111111111111} would give undefined results. What is worse, if
4112 a minimum quantifier for a parenthesized subpattern overflowed and became
4113 negative, the calculation of the memory size went wrong. This could have
4114 led to memory overwriting.
4116 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4118 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4119 operating environments where this matters.
4121 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4122 PCRE options from within the C++ wrapper.
4124 5. Named capturing subpatterns were not being correctly counted when a pattern
4125 was compiled. This caused two problems: (a) If there were more than 100
4126 such subpatterns, the calculation of the memory needed for the whole
4127 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4128 back references of the form \12, where the number was greater than 9, were
4129 not recognized as back references, even though there were sufficient
4130 previous subpatterns.
4132 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4133 versions of gcc, e.g. 2.95.4.
4136 Version 6.1 21-Jun-05
4137 ---------------------
4139 1. There was one reference to the variable "posix" in pcretest.c that was not
4140 surrounded by "#if !defined NOPOSIX".
4142 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4143 the cross-check on the old pcre_info() function, for the benefit of the
4144 cut-down version of PCRE that is currently imported into Exim.
4146 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4147 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4148 patterns that start (?i) but I don't think that matters. The use of (?i) is
4149 just an example; this all applies to the other options as well.
4151 4. Since libtool seems to echo the compile commands it is issuing, the output
4152 from "make" can be reduced a bit by putting "@" in front of each libtool
4153 compile command.
4155 5. Patch from the folks at Google for configure.in to be a bit more thorough
4156 in checking for a suitable C++ installation before trying to compile the
4157 C++ stuff. This should fix a reported problem when a compiler was present,
4158 but no suitable headers.
4160 6. The man pages all had just "PCRE" as their title. I have changed them to
4161 be the relevant file name. I have also arranged that these names are
4162 retained in the file doc/pcre.txt, which is a concatenation in text format
4163 of all the man pages except the little individual ones for each function.
4165 7. The NON-UNIX-USE file had not been updated for the different set of source
4166 files that come with release 6. I also added a few comments about the C++
4167 wrapper.
4170 Version 6.0 07-Jun-05
4171 ---------------------
4173 1. Some minor internal re-organization to help with my DFA experiments.
4175 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4176 didn't matter for the library itself when fully configured, but did matter
4177 when compiling without UCP support, or within Exim, where the ucp files are
4178 not imported.
4180 3. Refactoring of the library code to split up the various functions into
4181 different source modules. The addition of the new DFA matching code (see
4182 below) to a single monolithic source would have made it really too
4183 unwieldy, quite apart from causing all the code to be include in a
4184 statically linked application, when only some functions are used. This is
4185 relevant even without the DFA addition now that patterns can be compiled in
4186 one application and matched in another.
4188 The downside of splitting up is that there have to be some external
4189 functions and data tables that are used internally in different modules of
4190 the library but which are not part of the API. These have all had their
4191 names changed to start with "_pcre_" so that they are unlikely to clash
4192 with other external names.
4194 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4195 a different (DFA) algorithm. Although it is slower than the original
4196 function, it does have some advantages for certain types of matching
4197 problem.
4199 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4200 including restarting after a partial match.
4202 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4203 defined when compiling for Windows was sent to me. I have put it into the
4204 code, though I have no means of testing or verifying it.
4206 7. Added the pcre_refcount() auxiliary function.
4208 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4209 match before or at the first newline in the subject string. In pcretest,
4210 the /f option on a pattern can be used to set this.
4212 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4213 would behave wrongly. This has been present in PCRE since release 4.0.
4215 10. A number of changes to the pcregrep command:
4217 (a) Refactored how -x works; insert ^(...)$ instead of setting
4218 PCRE_ANCHORED and checking the length, in preparation for adding
4219 something similar for -w.
4221 (b) Added the -w (match as a word) option.
4223 (c) Refactored the way lines are read and buffered so as to have more
4224 than one at a time available.
4226 (d) Implemented a pcregrep test script.
4228 (e) Added the -M (multiline match) option. This allows patterns to match
4229 over several lines of the subject. The buffering ensures that at least
4230 8K, or the rest of the document (whichever is the shorter) is available
4231 for matching (and similarly the previous 8K for lookbehind assertions).
4233 (f) Changed the --help output so that it now says
4235 -w, --word-regex(p)
4237 instead of two lines, one with "regex" and the other with "regexp"
4238 because that confused at least one person since the short forms are the
4239 same. (This required a bit of code, as the output is generated
4240 automatically from a table. It wasn't just a text change.)
4242 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4243 option but starts with a hyphen. Could be a pattern or a path name
4244 starting with a hyphen, for instance.
4246 (h) "-" can be given as a file name to represent stdin.
4248 (i) When file names are being printed, "(standard input)" is used for
4249 the standard input, for compatibility with GNU grep. Previously
4250 "<stdin>" was used.
4252 (j) The option --label=xxx can be used to supply a name to be used for
4253 stdin when file names are being printed. There is no short form.
4255 (k) Re-factored the options decoding logic because we are going to add
4256 two more options that take data. Such options can now be given in four
4257 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4259 (l) Added the -A, -B, and -C options for requesting that lines of context
4260 around matches be printed.
4262 (m) Added the -L option to print the names of files that do not contain
4263 any matching lines, that is, the complement of -l.
4265 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4266 continue to scan other files.
4268 (o) The -s option was incorrectly implemented. For compatibility with other
4269 greps, it now suppresses the error message for a non-existent or non-
4270 accessible file (but not the return code). There is a new option called
4271 -q that suppresses the output of matching lines, which was what -s was
4272 previously doing.
4274 (p) Added --include and --exclude options to specify files for inclusion
4275 and exclusion when recursing.
4277 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4278 Hopefully, it now does.
4280 12. Missing cast in pcre_study().
4282 13. Added an "uninstall" target to the makefile.
4284 14. Replaced "extern" in the function prototypes in Makefile.in with
4285 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4286 world, but is set differently for Windows.
4288 15. Added a second compiling function called pcre_compile2(). The only
4289 difference is that it has an extra argument, which is a pointer to an
4290 integer error code. When there is a compile-time failure, this is set
4291 non-zero, in addition to the error test pointer being set to point to an
4292 error message. The new argument may be NULL if no error number is required
4293 (but then you may as well call pcre_compile(), which is now just a
4294 wrapper). This facility is provided because some applications need a
4295 numeric error indication, but it has also enabled me to tidy up the way
4296 compile-time errors are handled in the POSIX wrapper.
4298 16. Added VPATH=.libs to the makefile; this should help when building with one
4299 prefix path and installing with another. (Or so I'm told by someone who
4300 knows more about this stuff than I do.)
4302 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4303 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4304 match everything, including newlines. This is not POSIX-compatible, but
4305 somebody wanted the feature. From pcretest it can be activated by using
4306 both the P and the s flags.
4308 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4310 19. libpcre.pc was being incorrectly installed as executable.
4312 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4313 it now also looks for '\r' so that it will work unmodified on Windows.
4315 21. Added Google's contributed C++ wrapper to the distribution.
4317 22. Added some untidy missing memory free() calls in pcretest, to keep
4318 Electric Fence happy when testing.
4322 Version 5.0 13-Sep-04
4323 ---------------------
4325 1. Internal change: literal characters are no longer packed up into items
4326 containing multiple characters in a single byte-string. Each character
4327 is now matched using a separate opcode. However, there may be more than one
4328 byte in the character in UTF-8 mode.
4330 2. The pcre_callout_block structure has two new fields: pattern_position and
4331 next_item_length. These contain the offset in the pattern to the next match
4332 item, and its length, respectively.
4334 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4335 insertion of callouts before each pattern item. Added the /C option to
4336 pcretest to make use of this.
4338 4. On the advice of a Windows user, the lines
4340 #if defined(_WIN32) || defined(WIN32)
4341 _setmode( _fileno( stdout ), 0x8000 );
4342 #endif /* defined(_WIN32) || defined(WIN32) */
4344 have been added to the source of pcretest. This apparently does useful
4345 magic in relation to line terminators.
4347 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4348 for the benefit of those environments where the "b" makes a difference.
4350 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4351 to know about it. I have put a hack into configure.in that adds in code
4352 to set GCC=yes if CC=icc. This seems to end up at a point in the
4353 generated configure script that is early enough to affect the setting of
4354 compiler options, which is what is needed, but I have no means of testing
4355 whether it really works. (The user who reported this had patched the
4356 generated configure script, which of course I cannot do.)
4358 LATER: After change 22 below (new libtool files), the configure script
4359 seems to know about icc (and also ecc). Therefore, I have commented out
4360 this hack in configure.in.
4362 7. Added support for pkg-config (2 patches were sent in).
4364 8. Negated POSIX character classes that used a combination of internal tables
4365 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4366 [[:^ascii]]. Typically, they would match almost any characters. The other
4367 POSIX classes were not broken in this way.
4369 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4370 to find the match, as PCRE was deluded into thinking that the match had to
4371 start at the start point or following a newline. The same bug applied to
4372 patterns with negative forward assertions or any backward assertions
4373 preceding ".*" at the start, unless the pattern required a fixed first
4374 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4376 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4377 starting at the last subject character, bytes beyond the end of the subject
4378 string were read.
4380 11. Renamed the variable "class" as "classbits" to make life easier for C++
4381 users. (Previously there was a macro definition, but it apparently wasn't
4382 enough.)
4384 12. Added the new field "tables" to the extra data so that tables can be passed
4385 in at exec time, or the internal tables can be re-selected. This allows
4386 a compiled regex to be saved and re-used at a later time by a different
4387 program that might have everything at different addresses.
4389 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4390 -R library as well as a -L library.
4392 14. The debugging options of pcretest (-d on the command line or D on a
4393 pattern) showed incorrect output for anything following an extended class
4394 that contained multibyte characters and which was followed by a quantifier.
4396 15. Added optional support for general category Unicode character properties
4397 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4398 support. It adds about 90K to the size of the library. The meanings of the
4399 inbuilt class escapes such as \d and \s have NOT been changed.
4401 16. Updated pcredemo.c to include calls to free() to release the memory for the
4402 compiled pattern.
4404 17. The generated file chartables.c was being created in the source directory
4405 instead of in the building directory. This caused the build to fail if the
4406 source directory was different from the building directory, and was
4407 read-only.
4409 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4410 file. No doubt somebody will tell me if they don't make sense... Also added
4411 Dan Mooney's comments about building on OpenVMS.
4413 19. Added support for partial matching via the PCRE_PARTIAL option for
4414 pcre_exec() and the \P data escape in pcretest.
4416 20. Extended pcretest with 3 new pattern features:
4418 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4419 write the compiled pattern to the file whose name is "rest-of-line".
4420 This is a straight binary dump of the data, with the saved pointer to
4421 the character tables forced to be NULL. The study data, if any, is
4422 written too. After writing, pcretest reads a new pattern.
4424 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4425 compiled pattern from the given file. There must not be any
4426 occurrences of "<" in the file name (pretty unlikely); if there are,
4427 pcretest will instead treat the initial "<" as a pattern delimiter.
4428 After reading in the pattern, pcretest goes on to read data lines as
4429 usual.
4431 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4432 and 16-bit fields in a compiled pattern, to simulate a pattern that
4433 was compiled on a host of opposite endianness.
4435 21. The pcre-exec() function can now cope with patterns that were compiled on
4436 hosts of opposite endianness, with this restriction:
4438 As for any compiled expression that is saved and used later, the tables
4439 pointer field cannot be preserved; the extra_data field in the arguments
4440 to pcre_exec() should be used to pass in a tables address if a value
4441 other than the default internal tables were used at compile time.
4443 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4444 now diagnosed as an error. Previously, most of the time, a negative number
4445 would have been treated as zero, but if in addition "ovector" was passed as
4446 NULL, a crash could occur.
4448 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4449 new versions from the libtool 1.5 distribution (the last one is a copy of
4450 a file called libtool.m4). This seems to have fixed the need to patch
4451 "configure" to support Darwin 1.3 (which I used to do). However, I still
4452 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4453 workstation).
4455 24. Changed the PCRE licence to be the more standard "BSD" licence.
4458 Version 4.5 01-Dec-03
4459 ---------------------
4461 1. There has been some re-arrangement of the code for the match() function so
4462 that it can be compiled in a version that does not call itself recursively.
4463 Instead, it keeps those local variables that need separate instances for
4464 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4465 needs to "recurse". Keeping track of where control must go is done by means
4466 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4467 hide most of the details from the main code, and operates only if
4468 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4469 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4470 operating.
4472 To make it easier for callers to provide specially tailored get/free
4473 functions for this usage, two new functions, pcre_stack_malloc, and
4474 pcre_stack_free, are used. They are always called in strict stacking order,
4475 and the size of block requested is always the same.
4477 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4478 PCRE has been compiled to use the stack or the heap for recursion. The
4479 -C option of pcretest uses this to show which version is compiled.
4481 A new data escape \S, is added to pcretest; it causes the amounts of store
4482 obtained and freed by both kinds of malloc/free at match time to be added
4483 to the output.
4485 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4486 what's available on my current Linux desktop machine.
4488 3. When matching a UTF-8 string, the test for a valid string at the start has
4489 been extended. If start_offset is not zero, PCRE now checks that it points
4490 to a byte that is the start of a UTF-8 character. If not, it returns
4491 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4492 this is necessary because there may be backward assertions in the pattern.
4493 When matching the same subject several times, it may save resources to use
4494 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4496 4. The code for checking the validity of UTF-8 strings has been tightened so
4497 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4498 containing "overlong sequences".
4500 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4501 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4502 should have been "&", but it just so happened that all the cases this let
4503 through by mistake were picked up later in the function.
4505 6. I had used a variable called "isblank" - this is a C99 function, causing
4506 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4508 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4509 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4510 so the version is shown; (c) stop "make test" echoing "./RunTest".
4512 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4514 9. The prototype for memmove() for systems that don't have it was using
4515 size_t, but the inclusion of the header that defines size_t was later. I've
4516 moved the #includes for the C headers earlier to avoid this.
4518 10. Added some adjustments to the code to make it easier to compiler on certain
4519 special systems:
4521 (a) Some "const" qualifiers were missing.
4522 (b) Added the macro EXPORT before all exported functions; by default this
4523 is defined to be empty.
4524 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4525 that it reads its output file name as an argument instead of writing
4526 to the standard output and assuming this can be redirected.
4528 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4529 class containing characters with values greater than 255, PCRE compilation
4530 went into a loop.
4532 12. A recursive reference to a subpattern that was within another subpattern
4533 that had a minimum quantifier of zero caused PCRE to crash. For example,
4534 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4535 recursion. If the recursively-called subpattern itself had a zero repeat,
4536 that was OK.
4538 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4539 buffer into which it was copied (for escape processing) was still set at
4540 1024, so long lines caused crashes.
4542 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4543 "internal error: code overflow...". This applied to any character class
4544 that was followed by a possessive quantifier.
4546 15. Modified the Makefile to add libpcre.la as a prerequisite for
4547 libpcreposix.la because I was told this is needed for a parallel build to
4548 work.
4550 16. If a pattern that contained .* following optional items at the start was
4551 studied, the wrong optimizing data was generated, leading to matching
4552 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4553 matching string must start with a or b or c. The correct conclusion for
4554 this pattern is that a match can start with any character.
4557 Version 4.4 13-Aug-03
4558 ---------------------
4560 1. In UTF-8 mode, a character class containing characters with values between
4561 127 and 255 was not handled correctly if the compiled pattern was studied.
4562 In fixing this, I have also improved the studying algorithm for such
4563 classes (slightly).
4565 2. Three internal functions had redundant arguments passed to them. Removal
4566 might give a very teeny performance improvement.
4568 3. Documentation bug: the value of the capture_top field in a callout is *one
4569 more than* the number of the hightest numbered captured substring.
4571 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4572 in incorrectly linking with a previously installed version. They now link
4573 explicitly with libpcre.la.
4575 5. configure.in no longer needs to recognize Cygwin specially.
4577 6. A problem in pcre.in for Windows platforms is fixed.
4579 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4580 pcretest, it used to include the size of the study block as part of its
4581 output. Unfortunately, the structure contains a field that has a different
4582 size on different hardware architectures. This meant that the tests that
4583 showed this size failed. As the block is currently always of a fixed size,
4584 this information isn't actually particularly useful in pcretest output, so
4585 I have just removed it.
4587 8. Three pre-processor statements accidentally did not start in column 1.
4588 Sadly, there are *still* compilers around that complain, even though
4589 standard C has not required this for well over a decade. Sigh.
4591 9. In pcretest, the code for checking callouts passed small integers in the
4592 callout_data field, which is a void * field. However, some picky compilers
4593 complained about the casts involved for this on 64-bit systems. Now
4594 pcretest passes the address of the small integer instead, which should get
4595 rid of the warnings.
4597 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4598 both compile and run time, and gives an error if an invalid UTF-8 sequence
4599 is found. There is a option for disabling this check in cases where the
4600 string is known to be correct and/or the maximum performance is wanted.
4602 11. In response to a bug report, I changed one line in Makefile.in from
4604 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4605 to
4606 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4608 to look similar to other lines, but I have no way of telling whether this
4609 is the right thing to do, as I do not use Windows. No doubt I'll get told
4610 if it's wrong...
4613 Version 4.3 21-May-03
4614 ---------------------
4616 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4617 Makefile.
4619 2. Some refactoring to improve the quality of the code:
4621 (i) The utf8_table... variables are now declared "const".
4623 (ii) The code for \cx, which used the "case flipping" table to upper case
4624 lower case letters, now just substracts 32. This is ASCII-specific,
4625 but the whole concept of \cx is ASCII-specific, so it seems
4626 reasonable.
4628 (iii) PCRE was using its character types table to recognize decimal and
4629 hexadecimal digits in the pattern. This is silly, because it handles
4630 only 0-9, a-f, and A-F, but the character types table is locale-
4631 specific, which means strange things might happen. A private
4632 table is now used for this - though it costs 256 bytes, a table is
4633 much faster than multiple explicit tests. Of course, the standard
4634 character types table is still used for matching digits in subject
4635 strings against \d.
4637 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4638 ending in _t are). So I've renamed it as ESC_tee.
4640 3. The first argument for regexec() in the POSIX wrapper should have been
4641 defined as "const".
4643 4. Changed pcretest to use malloc() for its buffers so that they can be
4644 Electric Fenced for debugging.
4646 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4647 to read one or more bytes before the start of the subject string. Often this
4648 had no effect on PCRE's behaviour, but in some circumstances it could
4649 provoke a segmentation fault.
4651 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4652 to try to read one or more bytes before the start of the subject string.
4654 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4655 UTF-8 support could misbehave in various ways if the subject string
4656 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4657 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4658 back over UTF-8 characters.)
4661 Version 4.2 14-Apr-03
4662 ---------------------
4664 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4666 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4667 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4668 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4669 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4670 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4672 Note: automatic setting of the BUILD variables is not yet working
4673 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4674 compile-time but not at link-time
4675 [LINK]: use for linking executables only
4676 make different versions for Windows and non-Windows
4677 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4678 libraries
4679 [LINK_FOR_BUILD]: new variable
4680 [OBJEXT]: use throughout
4681 [EXEEXT]: use throughout
4682 <winshared>: new target
4683 <wininstall>: new target
4684 <dftables.o>: use native compiler
4685 <dftables>: use native linker
4686 <install>: handle Windows platform correctly
4687 <clean>: ditto
4688 <check>: ditto
4689 copy DLL to top builddir before testing
4691 As part of these changes, -no-undefined was removed again. This was reported
4692 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4693 in any case.
4695 3. Some tidies to get rid of compiler warnings:
4697 . In the match_data structure, match_limit was an unsigned long int, whereas
4698 match_call_count was an int. I've made them both unsigned long ints.
4700 . In pcretest the fact that a const uschar * doesn't automatically cast to
4701 a void * provoked a warning.
4703 . Turning on some more compiler warnings threw up some "shadow" variables
4704 and a few more missing casts.
4706 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4707 option, a class that contained a single character with a value between 128
4708 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4710 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4711 option, a class that contained several characters, but with at least one
4712 whose value was between 128 and 255 caused PCRE to crash.
4715 Version 4.1 12-Mar-03
4716 ---------------------
4718 1. Compiling with gcc -pedantic found a couple of places where casts were
4719 needed, and a string in dftables.c that was longer than standard compilers are
4720 required to support.
4722 2. Compiling with Sun's compiler found a few more places where the code could
4723 be tidied up in order to avoid warnings.
4725 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4726 first of these names is deprecated in the latest Autoconf in favour of the name
4727 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4728 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4729 analogy I have changed it to CFLAGS_FOR_BUILD.
4731 4. Added -no-undefined to the linking command in the Makefile, because this is
4732 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4733 linking step for the pcreposix library.
4735 5. PCRE was failing to diagnose the case of two named groups with the same
4736 name.
4738 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4739 literal character that is needed in the subject for a match, and scans along to
4740 ensure that it is present before embarking on the full matching process. This
4741 saves time in cases of nested unlimited repeats that are never going to match.
4742 Problem: the scan can take a lot of time if the subject is very long (e.g.
4743 megabytes), thus penalizing straightforward matches. It is now done only if the
4744 amount of subject to be scanned is less than 1000 bytes.
4746 7. A lesser problem with the same optimization is that it was recording the
4747 first character of an anchored pattern as "needed", thus provoking a search
4748 right along the subject, even when the first match of the pattern was going to
4749 fail. The "needed" character is now not set for anchored patterns, unless it
4750 follows something in the pattern that is of non-fixed length. Thus, it still
4751 fulfils its original purpose of finding quick non-matches in cases of nested
4752 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4755 Version 4.0 17-Feb-03
4756 ---------------------
4758 1. If a comment in an extended regex that started immediately after a meta-item
4759 extended to the end of string, PCRE compiled incorrect data. This could lead to
4760 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4762 2. Moved to autoconf 2.53 and libtool 1.4.2.
4764 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4765 the special perltest8 script is no longer needed - all the tests can be run
4766 from a single perltest script.
4768 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4769 by \s. It has now been removed in PCRE. This means it isn't recognized as
4770 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4771 class [:space:] *does* include VT, thereby creating a mess.
4773 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4774 space and tab.
4776 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4777 its new features into the main test script, reducing the number of scripts.
4779 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4780 were backward compatible, and made the (?i) apply to the whole pattern, as if
4781 /i were given. Now it behaves more logically, and applies the option setting
4782 only to what follows. PCRE has been changed to follow suit. However, if it
4783 finds options settings right at the start of the pattern, it extracts them into
4784 the global options, as before. Thus, they show up in the info data.
4786 8. Added support for the \Q...\E escape sequence. Characters in between are
4787 treated as literals. This is slightly different from Perl in that $ and @ are
4788 also handled as literals inside the quotes. In Perl, they will cause variable
4789 interpolation. Note the following examples:
4791 Pattern PCRE matches Perl matches
4793 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4794 \Qabc\$xyz\E abc\$xyz abc\$xyz
4795 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4797 For compatibility with Perl, \Q...\E sequences are recognized inside character
4798 classes as well as outside them.
4800 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4801 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4802 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4803 signed/unsigned warnings.
4805 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4806 option for pcretest, so I've replaced it by a simple function that does just
4807 that job.
4809 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4810 "pcregrep -".
4812 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4813 Java package. This provides some syntactic sugar for simple cases of what my
4814 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4815 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4816 item, you can use this simplified notation. Note that only makes sense with
4817 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4818 greediness, whatever the setting of the PCRE_UNGREEDY option.
4820 13. A change of greediness default within a pattern was not taking effect at
4821 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4822 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4823 was abstracted outside.
4825 14. PCRE now supports the \G assertion. It is true when the current matching
4826 position is at the start point of the match. This differs from \A when the
4827 starting offset is non-zero. Used with the /g option of pcretest (or similar
4828 code), it works in the same way as it does for Perl's /g option. If all
4829 alternatives of a regex begin with \G, the expression is anchored to the start
4830 match position, and the "anchored" flag is set in the compiled expression.
4832 15. Some bugs concerning the handling of certain option changes within patterns
4833 have been fixed. These applied to options other than (?ims). For example,
4834 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4835 been the other way round. Some of this was related to change 7 above.
4837 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4838 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4839 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4840 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4842 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4843 mode. Unlike ".", it always matches newline, whatever the setting of
4844 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4845 assertions. Perl allows it, but it doesn't (in general) work because it can't
4846 calculate the length of the lookbehind. At least, that's the case for Perl
4847 5.8.0 - I've been told they are going to document that it doesn't work in
4848 future.
4850 18. Added an error diagnosis for escapes that PCRE does not support: these are
4851 \L, \l, \N, \P, \p, \U, \u, and \X.
4853 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4854 reading past the end of the pattern in cases such as /[abcd/.
4856 20. PCRE was getting more memory than necessary for patterns with classes that
4857 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4859 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4860 compiling PCRE for use with Virtual Pascal.
4862 22. Small fix to the Makefile to make it work properly if the build is done
4863 outside the source tree.
4865 23. Added a new extension: a condition to go with recursion. If a conditional
4866 subpattern starts with (?(R) the "true" branch is used if recursion has
4867 happened, whereas the "false" branch is used only at the top level.
4869 24. When there was a very long string of literal characters (over 255 bytes
4870 without UTF support, over 250 bytes with UTF support), the computation of how
4871 much memory was required could be incorrect, leading to segfaults or other
4872 strange effects.
4874 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4875 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4876 there was a subsequent back reference to those brackets. This meant that, for
4877 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4878 possible to check for precisely this case. All we can do is abandon the
4879 optimization if .* occurs inside capturing brackets when there are any back
4880 references whatsoever. (See below for a better fix that came later.)
4882 26. The handling of the optimization for finding the first character of a
4883 non-anchored pattern, and for finding a character that is required later in the
4884 match were failing in some cases. This didn't break the matching; it just
4885 failed to optimize when it could. The way this is done has been re-implemented.
4887 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4889 28. Added a new feature that provides some of the functionality that Perl
4890 provides with (?{...}). The facility is termed a "callout". The way it is done
4891 in PCRE is for the caller to provide an optional function, by setting
4892 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4893 global variable. By default it is unset, which disables all calling out. To get
4894 the function called, the regex must include (?C) at appropriate points. This
4895 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4896 This provides a means of identifying different callout points. When PCRE
4897 reaches such a point in the regex, if pcre_callout has been set, the external
4898 function is called. It is provided with data in a structure called
4899 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4900 matching continues; if it returns a non-zero value, the match at the current
4901 point fails. However, backtracking will occur if possible. [This was changed
4902 later and other features added - see item 49 below.]
4904 29. pcretest is upgraded to test the callout functionality. It provides a
4905 callout function that displays information. By default, it shows the start of
4906 the match and the current position in the text. There are some new data escapes
4907 to vary what happens:
4909 \C+ in addition, show current contents of captured substrings
4910 \C- do not supply a callout function
4911 \C!n return 1 when callout number n is reached
4912 \C!n!m return 1 when callout number n is reached for the mth time
4914 30. If pcregrep was called with the -l option and just a single file name, it
4915 output "<stdin>" if a match was found, instead of the file name.
4917 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4918 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4919 pcre_exec(). This saves a malloc/free per call. The default value of
4920 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4921 when configuring.
4923 32. The default maximum size of a compiled pattern is 64K. There have been a
4924 few cases of people hitting this limit. The code now uses macros to handle the
4925 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4926 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4927 configuring. Tests 2 and 5 work only with 2-byte links because they output
4928 debugging information about compiled patterns.
4930 33. Internal code re-arrangements:
4932 (a) Moved the debugging function for printing out a compiled regex into
4933 its own source file (printint.c) and used #include to pull it into
4934 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4935 separate copies.
4937 (b) Defined the list of op-code names for debugging as a macro in
4938 internal.h so that it is next to the definition of the opcodes.
4940 (c) Defined a table of op-code lengths for simpler skipping along compiled
4941 code. This is again a macro in internal.h so that it is next to the
4942 definition of the opcodes.
4944 34. Added support for recursive calls to individual subpatterns, along the
4945 lines of Robin Houston's patch (but implemented somewhat differently).
4947 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4948 allow it to read and process whole directories in Win32. This code was
4949 contributed by Lionel Fourquaux; it has not been tested by me.
4951 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4952 used to name a group. Names consist of alphanumerics and underscores, and must
4953 be unique. Back references use the syntax (?P=name) and recursive calls use
4954 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4955 numbers. The function pcre_fullinfo() can be used after compilation to extract
4956 a name/number map. There are three relevant calls:
4958 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4959 PCRE_INFO_NAMECOUNT yields the number of entries
4960 PCRE_INFO_NAMETABLE yields a pointer to the map.
4962 The map is a vector of fixed-size entries. The size of each entry depends on
4963 the length of the longest name used. The first two bytes of each entry are the
4964 group number, most significant byte first. There follows the corresponding
4965 name, zero terminated. The names are in alphabetical order.
4967 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4968 case instead of 255. Making it the same both with and without UTF-8 support
4969 means that the same test output works with both.
4971 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4972 calling malloc() with a zero argument.
4974 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4975 optimization. I've improved things by keeping a bitmap of backreferences with
4976 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4977 fact referenced, the optimization can be applied. It is unlikely that a
4978 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4979 the match to follow \n) will appear inside brackets with a number greater than
4980 31, but if it does, any back reference > 31 suppresses the optimization.
4982 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4983 of disabling numbered capturing parentheses. Any opening parenthesis that is
4984 not followed by ? behaves as if it were followed by ?: but named parentheses
4985 can still be used for capturing (and they will acquire numbers in the usual
4986 way).
4988 41. Redesigned the return codes from the match() function into yes/no/error so
4989 that errors can be passed back from deep inside the nested calls. A malloc
4990 failure while inside a recursive subpattern call now causes the
4991 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4993 42. It is now possible to set a limit on the number of times the match()
4994 function is called in a call to pcre_exec(). This facility makes it possible to
4995 limit the amount of recursion and backtracking, though not in a directly
4996 obvious way, because the match() function is used in a number of different
4997 circumstances. The count starts from zero for each position in the subject
4998 string (for non-anchored patterns). The default limit is, for compatibility, a
4999 large number, namely 10 000 000. You can change this in two ways:
5001 (a) When configuring PCRE before making, you can use --with-match-limit=n
5002 to set a default value for the compiled library.
5004 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5005 a different value is set. See 45 below.
5007 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5009 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5010 of things that can be changed at compile time. The first argument specifies
5011 what is wanted and the second points to where the information is to be placed.
5012 The current list of available information is:
5016 The output is an integer that is set to one if UTF-8 support is available;
5017 otherwise it is set to zero.
5021 The output is an integer that it set to the value of the code that is used for
5022 newline. It is either LF (10) or CR (13).
5026 The output is an integer that contains the number of bytes used for internal
5027 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5031 The output is an integer that contains the threshold above which the POSIX
5032 interface uses malloc() for output vectors. See item 31 above.
5036 The output is an unsigned integer that contains the default limit of the number
5037 of match() calls in a pcre_exec() execution. See 42 above.
5039 44. pcretest has been upgraded by the addition of the -C option. This causes it
5040 to extract all the available output from the new pcre_config() function, and to
5041 output it. The program then exits immediately.
5043 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5044 order to support additional features. One way would have been to define
5045 pcre_exec2() (for example) with extra arguments, but this would not have been
5046 extensible, and would also have required all calls to the original function to
5047 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5048 is used for passing in "extra" data from pcre_study().
5050 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5051 contains the following fields:
5053 flags a bitmap indicating which of the following fields are set
5054 study_data opaque data from pcre_study()
5055 match_limit a way of specifying a limit on match() calls for a specific
5056 call to pcre_exec()
5057 callout_data data for callouts (see 49 below)
5059 The flag bits are also defined in pcre.h, and are
5065 The pcre_study() function now returns one of these new pcre_extra blocks, with
5066 the actual study data pointed to by the study_data field, and the
5067 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5068 before. That is, this change is entirely upwards-compatible and requires no
5069 change to existing code.
5071 If you want to pass in additional data to pcre_exec(), you can either place it
5072 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5073 block.
5075 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5076 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5077 times with different match limits, until it finds the minimum value needed for
5078 pcre_exec() to complete. The value is then output. This can be instructive; for
5079 most simple matches the number is quite small, but for pathological cases it
5080 gets very large very quickly.
5082 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5083 returns the size of the data block pointed to by the study_data field in a
5084 pcre_extra block, that is, the value that was passed as the argument to
5085 pcre_malloc() when PCRE was getting memory in which to place the information
5086 created by pcre_study(). The fourth argument should point to a size_t variable.
5087 pcretest has been extended so that this information is shown after a successful
5088 pcre_study() call when information about the compiled regex is being displayed.
5090 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5091 because what follows is always an absolute path. (Later: it turns out that this
5092 is more than cosmetic for MinGW, because it doesn't like empty path
5093 components.)
5095 49. Some changes have been made to the callout feature (see 28 above):
5097 (i) A callout function now has three choices for what it returns:
5099 0 => success, carry on matching
5100 > 0 => failure at this point, but backtrack if possible
5101 < 0 => serious error, return this value from pcre_exec()
5103 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5104 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5105 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5106 use by callout functions. It will never be used by PCRE itself.
5108 (ii) The pcre_extra structure (see 45 above) has a void * field called
5109 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5110 pcre_callout_block structure has a field of the same name. The contents of
5111 the field passed in the pcre_extra structure are passed to the callout
5112 function in the corresponding field in the callout block. This makes it
5113 easier to use the same callout-containing regex from multiple threads. For
5114 testing, the pcretest program has a new data escape
5116 \C*n pass the number n (may be negative) as callout_data
5118 If the callout function in pcretest receives a non-zero value as
5119 callout_data, it returns that value.
5121 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5122 there were some redundant $(CFLAGS) in commands that are now specified as
5123 $(LINK), which already includes $(CFLAGS).
5125 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5126 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5127 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5128 one-byte characters throughout. Note that case-insensitive matching applies
5129 only to characters whose values are less than 256. PCRE doesn't support the
5130 notion of cases for higher-valued characters.
5132 (i) A character class whose characters are all within 0-255 is handled as
5133 a bit map, and the map is inverted for negative classes. Previously, a
5134 character > 255 always failed to match such a class; however it should
5135 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5137 (ii) A negated character class with a single character < 255 is coded as
5138 "not this character" (OP_NOT). This wasn't working properly when the test
5139 character was multibyte, either singly or repeated.
5141 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5142 mode, for example: \x{100}{2,3}.
5144 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5145 singly or repeated) now correctly test multibyte characters. However,
5146 PCRE doesn't recognize any characters with values greater than 255 as
5147 digits, spaces, or word characters. Such characters always match \D, \S,
5148 and \W, and never match \d, \s, or \w.
5150 (v) Classes may now contain characters and character ranges with values
5151 greater than 255. For example: [ab\x{100}-\x{400}].
5153 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5154 PCRE in UTF-8 mode.
5156 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5157 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5158 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5159 value.)
5161 53. The single man page has become too large. I have therefore split it up into
5162 a number of separate man pages. These also give rise to individual HTML pages;
5163 these are now put in a separate directory, and there is an index.html page that
5164 lists them all. Some hyperlinking between the pages has been installed.
5166 54. Added convenience functions for handling named capturing parentheses.
5168 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5169 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5170 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5171 are faulted.
5173 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5174 calling configure. These values are used when compiling the dftables.c program
5175 which is run to generate the source of the default character tables. They
5176 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5177 you will need to set these values.
5179 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5182 Version 3.9 02-Jan-02
5183 ---------------------
5185 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5187 2. If --disable-static was given, the building process failed when trying to
5188 build pcretest and pcregrep. (For some reason it was using libtool to compile
5189 them, which is not right, as they aren't part of the library.)
5192 Version 3.8 18-Dec-01
5193 ---------------------
5195 1. The experimental UTF-8 code was completely screwed up. It was packing the
5196 bytes in the wrong order. How dumb can you get?
5199 Version 3.7 29-Oct-01
5200 ---------------------
5202 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5203 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5204 this didn't happen under Solaris 8, where I normally test things.
5206 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5207 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5208 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5209 this fix an hour or so after the initial 3.7 release.)
5212 Version 3.6 23-Oct-01
5213 ---------------------
5215 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5216 offsets passed as NULL with zero offset count.
5218 2. The config.guess and config.sub files had not been updated when I moved to
5219 the latest autoconf.
5222 Version 3.5 15-Aug-01
5223 ---------------------
5225 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5226 had been forgotten.
5228 2. By using declared but undefined structures, we can avoid using "void"
5229 definitions in pcre.h while keeping the internal definitions of the structures
5230 private.
5232 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5233 user point of view, this means that both static and shared libraries are built
5234 by default, but this can be individually controlled. More of the work of
5235 handling this static/shared cases is now inside libtool instead of PCRE's make
5236 file.
5238 4. The pcretest utility is now installed along with pcregrep because it is
5239 useful for users (to test regexs) and by doing this, it automatically gets
5240 relinked by libtool. The documentation has been turned into a man page, so
5241 there are now .1, .txt, and .html versions in /doc.
5243 5. Upgrades to pcregrep:
5244 (i) Added long-form option names like gnu grep.
5245 (ii) Added --help to list all options with an explanatory phrase.
5246 (iii) Added -r, --recursive to recurse into sub-directories.
5247 (iv) Added -f, --file to read patterns from a file.
5249 6. pcre_exec() was referring to its "code" argument before testing that
5250 argument for NULL (and giving an error if it was NULL).
5252 7. Upgraded Makefile.in to allow for compiling in a different directory from
5253 the source directory.
5255 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5256 options bits, the pointer it was passed was to an int instead of to an unsigned
5257 long int. This mattered only on 64-bit systems.
5259 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5260 generated) instead of pcre.in, which it its source. Also made the same change
5261 in several of the .c files.
5263 10. A new release of gcc defines printf() as a macro, which broke pcretest
5264 because it had an ifdef in the middle of a string argument for printf(). Fixed
5265 by using separate calls to printf().
5267 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5268 script, to force use of CR or LF instead of \n in the source. On non-Unix
5269 systems, the value can be set in config.h.
5271 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5272 absolute limit. Changed the text of the error message to make this clear, and
5273 likewise updated the man page.
5275 13. The limit of 99 on the number of capturing subpatterns has been removed.
5276 The new limit is 65535, which I hope will not be a "real" limit.
5279 Version 3.4 22-Aug-00
5280 ---------------------
5282 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5284 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5287 Version 3.3 01-Aug-00
5288 ---------------------
5290 1. If an octal character was given, but the value was greater than \377, it
5291 was not getting masked to the least significant bits, as documented. This could
5292 lead to crashes in some systems.
5294 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5295 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5297 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5298 These just pass their arguments on to (pcre_free)(), but they are provided
5299 because some uses of PCRE bind it to non-C systems that can call its functions,
5300 but cannot call free() or pcre_free() directly.
5302 4. Add "make test" as a synonym for "make check". Corrected some comments in
5303 the Makefile.
5305 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5306 Makefile.
5308 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5309 command called pgrep for grepping around the active processes.
5311 7. Added the beginnings of support for UTF-8 character strings.
5313 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5314 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5315 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5316 out for the ar command.)
5319 Version 3.2 12-May-00
5320 ---------------------
5322 This is purely a bug fixing release.
5324 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5325 of ZA. This was just one example of several cases that could provoke this bug,
5326 which was introduced by change 9 of version 2.00. The code for breaking
5327 infinite loops after an iteration that matches an empty string was't working
5328 correctly.
5330 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5331 when matched against abbab (for example). After matching an empty string, it
5332 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5333 caused it to match further down the string than it should.
5335 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5336 was there because it doesn't seem to be needed, and it causes trouble on some
5337 systems, as it is not a Standard C header. It has been removed.
5339 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5340 were reported on the Macintosh. The changes were from
5342 while ((c = *(++ptr)) != 0 && c != '\n');
5343 to
5344 while ((c = *(++ptr)) != 0 && c != '\n') ;
5346 Totally extraordinary, but if that's what it takes...
5348 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5349 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5350 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5351 assumes the way PCRE uses memmove() (always moving upwards).
5353 6. PCRE is being used in one environment where strchr() is not available. There
5354 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5355 faster code anyway.
5358 Version 3.1 09-Feb-00
5359 ---------------------
5361 The only change in this release is the fixing of some bugs in Makefile.in for
5362 the "install" target:
5364 (1) It was failing to install pcreposix.h.
5366 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5369 Version 3.0 01-Feb-00
5370 ---------------------
5372 1. Add support for the /+ modifier to perltest (to output $` like it does in
5373 pcretest).
5375 2. Add support for the /g modifier to perltest.
5377 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5378 matches null strings.
5380 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5381 pattern. Perl treats empty patterns specially - it reuses the most recent
5382 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5383 effect.
5385 5. The POSIX interface was broken in that it was just handing over the POSIX
5386 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5387 required a bigger vector, with some working space on the end. This means that
5388 the POSIX wrapper now has to get and free some memory, and copy the results.
5390 6. Added some simple autoconf support, placing the test data and the
5391 documentation in separate directories, re-organizing some of the
5392 information files, and making it build pcre-config (a GNU standard). Also added
5393 libtool support for building PCRE as a shared library, which is now the
5394 default.
5396 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5397 09 are not valid octal constants. Single digits will be used for minor values
5398 less than 10.
5400 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5401 existing programs that set these in the POSIX interface can use PCRE without
5402 modification.
5404 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5405 return all that pcre_info() returns, plus additional data. The pcre_info()
5406 function is retained for compatibility, but is considered to be obsolete.
5408 10. Added experimental recursion feature (?R) to handle one common case that
5409 Perl 5.6 will be able to do with (?p{...}).
5411 11. Added support for POSIX character classes like [:alpha:], which Perl is
5412 adopting.
5415 Version 2.08 31-Aug-99
5416 ----------------------
5418 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5419 trying to match at the startoffset position, but instead was moving forward to
5420 the next newline as if a previous match had failed.
5422 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5423 and could get into a loop if a null string was matched other than at the start
5424 of the subject.
5426 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5427 be distinguished at compile time, and for completeness also added PCRE_DATE.
5429 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5430 in GnuWin32 environments.
5433 Version 2.07 29-Jul-99
5434 ----------------------
5436 1. The documentation is now supplied in plain text form and HTML as well as in
5437 the form of man page sources.
5439 2. C++ compilers don't like assigning (void *) values to other pointer types.
5440 In particular this affects malloc(). Although there is no problem in Standard
5441 C, I've put in casts to keep C++ compilers happy.
5443 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5444 should be (const char *).
5446 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5447 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5448 However, I haven't made this a standard facility. The documentation doesn't
5449 mention it, and the Makefile doesn't support it.
5451 5. The Makefile now contains an "install" target, with editable destinations at
5452 the top of the file. The pcretest program is not installed.
5454 6. pgrep -V now gives the PCRE version number and date.
5456 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5457 causing the entire string to be ignored, instead of just the last character.
5459 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5460 non-matching string, it can take a very, very long time, even for strings of
5461 quite modest length, because of the nested recursion. PCRE now does better in
5462 some of these cases. It does this by remembering the last required literal
5463 character in the pattern, and pre-searching the subject to ensure it is present
5464 before running the real match. In other words, it applies a heuristic to detect
5465 some types of certain failure quickly, and in the above example, if presented
5466 with a string that has no trailing " it gives "no match" very quickly.
5468 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5469 other alternatives are tried instead.
5472 Version 2.06 09-Jun-99
5473 ----------------------
5475 1. Change pcretest's output for amount of store used to show just the code
5476 space, because the remainder (the data block) varies in size between 32-bit and
5477 64-bit systems.
5479 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5480 start matching at. This allows lookbehinds to work when searching for multiple
5481 occurrences in a string.
5483 3. Added additional options to pcretest for testing multiple occurrences:
5485 /+ outputs the rest of the string that follows a match
5486 /g loops for multiple occurrences, using the new startoffset argument
5487 /G loops for multiple occurrences by passing an incremented pointer
5489 4. PCRE wasn't doing the "first character" optimization for patterns starting
5490 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5491 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5492 the letter 'x'. On long subject strings, this gives a significant speed-up.
5495 Version 2.05 21-Apr-99
5496 ----------------------
5498 1. Changed the type of magic_number from int to long int so that it works
5499 properly on 16-bit systems.
5501 2. Fixed a bug which caused patterns starting with .* not to work correctly
5502 when the subject string contained newline characters. PCRE was assuming
5503 anchoring for such patterns in all cases, which is not correct because .* will
5504 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5505 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5506 must be retried after every newline in the subject.
5509 Version 2.04 18-Feb-99
5510 ----------------------
5512 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5513 computation of the store needed to hold the pattern was incorrect (too large).
5514 If such patterns were nested a few deep, this could multiply and become a real
5515 problem.
5517 2. Added /M option to pcretest to show the memory requirement of a specific
5518 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5520 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5521 compiled in such a way that the backtracking after subsequent failure was
5522 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5523 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5526 Version 2.03 02-Feb-99
5527 ----------------------
5529 1. Fixed typo and small mistake in man page.
5531 2. Added 4th condition (GPL supersedes if conflict) and created separate
5532 LICENCE file containing the conditions.
5534 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5535 Perl, that is the internal \ allows the delimiter to be included in the
5536 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5537 the final delimiter, add \ to the end of the pattern (to test the error).
5539 4. Added the convenience functions for extracting substrings after a successful
5540 match. Updated pcretest to make it able to test these functions.
5543 Version 2.02 14-Jan-99
5544 ----------------------
5546 1. Initialized the working variables associated with each extraction so that
5547 their saving and restoring doesn't refer to uninitialized store.
5549 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5550 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5551 fix the problem.
5553 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5554 calls, and wasn't printing the correct value for compiling calls. Increased the
5555 default value of LOOPREPEAT, and the number of significant figures in the
5556 times.
5558 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5560 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5561 a building problem on Windows NT with a FAT file system.
5564 Version 2.01 21-Oct-98
5565 ----------------------
5567 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5568 to character tables built by pcre_maketables() in the current locale. If NULL
5569 is passed, the default tables are used.
5572 Version 2.00 24-Sep-98
5573 ----------------------
5575 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5576 it any more.
5578 2. Allow quantification of (?>) groups, and make it work correctly.
5580 3. The first character computation wasn't working for (?>) groups.
5582 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5583 end of the subject) and add 5.005's \z, which really does match only at the
5584 very end of the subject.
5586 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5588 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5589 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5590 localized options. All options to pcre_study() were also removed.
5592 7. Add other new features from 5.005:
5594 $(?<= positive lookbehind
5595 $(?<! negative lookbehind
5596 (?imsx-imsx) added the unsetting capability
5597 such a setting is global if at outer level; local otherwise
5598 (?imsx-imsx:) non-capturing groups with option setting
5599 (?(cond)re|re) conditional pattern matching
5601 A backreference to itself in a repeated group matches the previous
5602 captured string.
5604 8. General tidying up of studying (both automatic and via "study")
5605 consequential on the addition of new assertions.
5607 9. As in 5.005, unlimited repeated groups that could match an empty substring
5608 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5609 runtime if any iteration does actually match an empty substring.
5611 10. Include the RunTest script in the distribution.
5613 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5614 discrepancies, some of which were old and were also with respect to 5.004. They
5615 have now been fixed.
5618 Version 1.09 28-Apr-98
5619 ----------------------
5621 1. A negated single character class followed by a quantifier with a minimum
5622 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5623 program crashes, or just wrong answers. This did not apply to negated classes
5624 containing more than one character, or to minima other than one.
5627 Version 1.08 27-Mar-98
5628 ----------------------
5630 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5632 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5633 latter must appear before anything that relies on it in the pattern.
5636 Version 1.07 16-Feb-98
5637 ----------------------
5639 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5640 repeat of a potentially empty string).
5643 Version 1.06 23-Jan-98
5644 ----------------------
5646 1. Added Markus Oberhumer's little patches for C++.
5648 2. Literal strings longer than 255 characters were broken.
5651 Version 1.05 23-Dec-97
5652 ----------------------
5654 1. Negated character classes containing more than one character were failing if
5655 PCRE_CASELESS was set at run time.
5658 Version 1.04 19-Dec-97
5659 ----------------------
5661 1. Corrected the man page, where some "const" qualifiers had been omitted.
5663 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5664 input syntax.
5666 3. Fixed memory leak which occurred when a regex with back references was
5667 matched with an offsets vector that wasn't big enough. The temporary memory
5668 that is used in this case wasn't being freed if the match failed.
5670 4. Tidied pcretest to ensure it frees memory that it gets.
5672 5. Temporary memory was being obtained in the case where the passed offsets
5673 vector was exactly big enough.
5675 6. Corrected definition of offsetof() from change 5 below.
5677 7. I had screwed up change 6 below and broken the rules for the use of
5678 setjmp(). Now fixed.
5681 Version 1.03 18-Dec-97
5682 ----------------------
5684 1. A erroneous regex with a missing opening parenthesis was correctly
5685 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5686 on some systems.
5688 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5689 it was reported that one broken compiler failed on the former because "code" is
5690 also an independent variable.
5692 3. The erroneous regex a[]b caused an array overrun reference.
5694 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5695 fail on data ending with that character. (It was going on too far, and checking
5696 the next character, typically a binary zero.) This was specific to the
5697 optimized code for single-character negative classes.
5699 5. Added a contributed patch from the TIN world which does the following:
5701 + Add an undef for memmove, in case the the system defines a macro for it.
5703 + Add a definition of offsetof(), in case there isn't one. (I don't know
5704 the reason behind this - offsetof() is part of the ANSI standard - but
5705 it does no harm).
5707 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5708 most of the places where whitespace preceded '#'. I have given up and
5709 allowed the remaining 2 cases to be at the margin.
5711 + Rename some variables in pcre to eliminate shadowing. This seems very
5712 pedantic, but does no harm, of course.
5714 6. Moved the call to setjmp() into its own function, to get rid of warnings
5715 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5717 7. Constructs such as \d{8,} were compiling into the equivalent of
5718 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5719 outcome, but in this particular case used more store than had been allocated,
5720 which caused the bug to be discovered because it threw up an internal error.
5722 8. The debugging code in both pcre and pcretest for outputting the compiled
5723 form of a regex was going wrong in the case of back references followed by
5724 curly-bracketed repeats.
5727 Version 1.02 12-Dec-97
5728 ----------------------
5730 1. Typos in pcre.3 and comments in the source fixed.
5732 2. Applied a contributed patch to get rid of places where it used to remove
5733 'const' from variables, and fixed some signed/unsigned and uninitialized
5734 variable warnings.
5736 3. Added the "runtest" target to Makefile.
5738 4. Set default compiler flag to -O2 rather than just -O.
5741 Version 1.01 19-Nov-97
5742 ----------------------
5744 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5745 like /([ab]*)*/, that is, for classes with more than one character in them.
5747 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5748 as /((?>a*))*/ (a PCRE_EXTRA facility).
5751 Version 1.00 18-Nov-97
5752 ----------------------
5754 1. Added compile-time macros to support systems such as SunOS4 which don't have
5755 memmove() or strerror() but have other things that can be used instead.
5757 2. Arranged that "make clean" removes the executables.
5760 Version 0.99 27-Oct-97
5761 ----------------------
5763 1. Fixed bug in code for optimizing classes with only one character. It was
5764 initializing a 32-byte map regardless, which could cause it to run off the end
5765 of the memory it had got.
5767 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5770 Version 0.98 22-Oct-97
5771 ----------------------
5773 1. Fixed bug in code for handling temporary memory usage when there are more
5774 back references than supplied space in the ovector. This could cause segfaults.
5777 Version 0.97 21-Oct-97
5778 ----------------------
5780 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5782 2. Optimized negated single characters not to use a bit map.
5784 3. Brought error texts together as macro definitions; clarified some of them;
5785 fixed one that was wrong - it said "range out of order" when it meant "invalid
5786 escape sequence".
5788 4. Changed some char * arguments to const char *.
5790 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5792 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5793 pcretest.
5796 Version 0.96 16-Oct-97
5797 ----------------------
5799 1. Added a simple "pgrep" utility to the distribution.
5801 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5802 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5803 where "ddd" means "one or more decimal digits".
5805 3. Fixed serious bug. If a pattern had a back reference, but the call to
5806 pcre_exec() didn't supply a large enough ovector to record the related
5807 identifying subpattern, the match always failed. PCRE now remembers the number
5808 of the largest back reference, and gets some temporary memory in which to save
5809 the offsets during matching if necessary, in order to ensure that
5810 backreferences always work.
5812 4. Increased the compatibility with Perl in a number of ways:
5814 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5815 to request this handling. The option can be set at compile or exec time.
5817 (b) $ matches before a terminating newline by default; an option
5818 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5819 mode). The option can be set at compile or exec time.
5821 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5822 the same as Perl's. If the decimal number it represents is less than 10
5823 or there aren't that many previous left capturing parentheses, an octal
5824 escape is read. Inside a character class, it's always an octal escape,
5825 even if it is a single digit.
5827 (d) An escaped but undefined alphabetic character is taken as a literal,
5828 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5829 escapes.
5831 (e) {0} is now permitted. (The previous item is removed from the compiled
5832 pattern).
5834 5. Changed all the names of code files so that the basic parts are no longer
5835 than 10 characters, and abolished the teeny "globals.c" file.
5837 6. Changed the handling of character classes; they are now done with a 32-byte
5838 bit map always.
5840 7. Added the -d and /D options to pcretest to make it possible to look at the
5841 internals of compilation without having to recompile pcre.
5844 Version 0.95 23-Sep-97
5845 ----------------------
5847 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5848 \x20 at the start of a run of normal characters. These were being treated as
5849 real characters, instead of the source characters being re-checked.
5852 Version 0.94 18-Sep-97
5853 ----------------------
5855 1. The functions are now thread-safe, with the caveat that the global variables
5856 containing pointers to malloc() and free() or alternative functions are the
5857 same for all threads.
5859 2. Get pcre_study() to generate a bitmap of initial characters for non-
5860 anchored patterns when this is possible, and use it if passed to pcre_exec().
5863 Version 0.93 15-Sep-97
5864 ----------------------
5866 1. /(b)|(:+)/ was computing an incorrect first character.
5868 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5869 but not actually doing anything yet.
5871 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5872 as Perl does (e.g. [-az] or [az-]).
5874 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5875 all possible positions.
5877 5. Split up into different modules to avoid including unneeded functions in a
5878 compiled binary. However, compile and exec are still in one module. The "study"
5879 function is split off.
5881 6. The character tables are now in a separate module whose source is generated
5882 by an auxiliary program - but can then be edited by hand if required. There are
5883 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5884 toupper() in the code.
5886 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5887 make them global. Abolish the function for setting them, as the caller can now
5888 set them directly.
5891 Version 0.92 11-Sep-97
5892 ----------------------
5894 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5895 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5897 2. Caseless matching was not working in character classes if the characters in
5898 the pattern were in upper case.
5900 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5902 4. Make PCRE_ANCHORED public and accept as a compile option.
5904 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5905 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5906 pass them.
5908 6. Give an error if bad option bits passed at compile or run time.
5910 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5911 pcretest to cause it to pass that flag.
5913 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5914 options, and the first character, if set.
5916 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5919 Version 0.91 10-Sep-97
5920 ----------------------
5922 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5923 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5925 2. PCRE was looping on encountering an indefinitely repeated back reference to
5926 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5927 Perl does - treats the match as successful.
5929 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5