ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1566 - (show annotations)
Tue Jun 9 17:45:25 2015 UTC (4 years, 5 months ago) by ph10
File size: 279840 byte(s)
Error occurred while calculating annotation data.
Fix group empty match 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 8.38 xx-xxx-xxxx
8 ------------------------
10 1. If a group that contained a recursive back reference also contained a
11 forward reference subroutine call followed by a non-forward-reference
12 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13 compile correct code, leading to undefined behaviour or an internally
14 detected error. This bug was discovered by the LLVM fuzzer.
16 2. Quantification of certain items (e.g. atomic back references) could cause
17 incorrect code to be compiled when recursive forward references were
18 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19 This bug was discovered by the LLVM fuzzer.
21 3. A repeated conditional group whose condition was a reference by name caused
22 a buffer overflow if there was more than one group with the given name.
23 This bug was discovered by the LLVM fuzzer.
25 4. A recursive back reference by name within a group that had the same name as
26 another group caused a buffer overflow. For example:
27 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
29 5. A forward reference by name to a group whose number is the same as the
30 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31 a buffer overflow at compile time. This bug was discovered by the LLVM
32 fuzzer.
34 6. A lookbehind assertion within a set of mutually recursive subpatterns could
35 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
37 7. Another buffer overflow bug involved duplicate named groups with a
38 reference between their definition, with a group that reset capture
39 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40 fixed by always allowing for more memory, even if not needed. (A proper fix
41 is implemented in PCRE2, but it involves more refactoring.)
43 8. There was no check for integer overflow in subroutine calls such as (?123).
45 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
46 being treated as a literal 'l' instead of causing an error.
48 10. There was a buffer overflow if pcre_exec() was called with an ovector of
49 size 1. This bug was found by american fuzzy lop.
51 11. If a non-capturing group containing a conditional group that could match
52 an empty string was repeated, it was not identified as matching an empty
53 string itself. For example: /^(?:(?(1)x|)+)+$()/.
56 Version 8.37 28-April-2015
57 --------------------------
59 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
60 for those parentheses to be closed with whatever has been captured so far.
61 However, it was failing to mark any other groups between the hightest
62 capture so far and the currrent group as "unset". Thus, the ovector for
63 those groups contained whatever was previously there. An example is the
64 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
66 2. If an assertion condition was quantified with a minimum of zero (an odd
67 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
69 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
70 unrecognized modifier, a crash could occur.
72 4. An attempt to do global matching in pcretest with a zero-length ovector
73 caused a crash.
75 5. Fixed a memory leak during matching that could occur for a subpattern
76 subroutine call (recursive or otherwise) if the number of captured groups
77 that had to be saved was greater than ten.
79 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
80 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
81 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
83 7. A UTF pattern containing a "not" match of a non-ASCII character and a
84 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
86 8. When a pattern is compiled, it remembers the highest back reference so that
87 when matching, if the ovector is too small, extra memory can be obtained to
88 use instead. A conditional subpattern whose condition is a check on a
89 capture having happened, such as, for example in the pattern
90 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
91 setting the highest backreference number. This mattered only if pcre_exec()
92 was called with an ovector that was too small to hold the capture, and there
93 was no other kind of back reference (a situation which is probably quite
94 rare). The effect of the bug was that the condition was always treated as
95 FALSE when the capture could not be consulted, leading to a incorrect
96 behaviour by pcre_exec(). This bug has been fixed.
98 9. A reference to a duplicated named group (either a back reference or a test
99 for being set in a conditional) that occurred in a part of the pattern where
100 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
101 to be incorrectly calculated, leading to overwriting.
103 10. A mutually recursive set of back references such as (\2)(\1) caused a
104 segfault at study time (while trying to find the minimum matching length).
105 The infinite loop is now broken (with the minimum length unset, that is,
106 zero).
108 11. If an assertion that was used as a condition was quantified with a minimum
109 of zero, matching went wrong. In particular, if the whole group had
110 unlimited repetition and could match an empty string, a segfault was
111 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
112 assertions to be quantified, but not if they are being used as conditions,
113 so the above pattern is faulted by Perl. PCRE has now been changed so that
114 it also rejects such patterns.
116 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
117 failed to allow the zero-repeat case if pcre2_exec() was called with an
118 ovector too small to capture the group.
120 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
121 Red Hat Product Security:
123 (a) A crash if /K and /F were both set with the option to save the compiled
124 pattern.
126 (b) Another crash if the option to print captured substrings in a callout
127 was combined with setting a null ovector, for example \O\C+ as a subject
128 string.
130 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
131 forward reference repeated a large (but limited) number of times within a
132 repeated outer group that has a zero minimum quantifier, caused incorrect
133 code to be compiled, leading to the error "internal error:
134 previously-checked referenced subpattern not found" when an incorrect
135 memory address was read. This bug was reported as "heap overflow",
136 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
137 CVE-2015-2325.
139 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
140 call within a group that also contained a recursive back reference caused
141 incorrect code to be compiled. This bug was reported as "heap overflow",
142 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
143 number CVE-2015-2326.
145 24. Computing the size of the JIT read-only data in advance has been a source
146 of various issues, and new ones are still appear unfortunately. To fix
147 existing and future issues, size computation is eliminated from the code,
148 and replaced by on-demand memory allocation.
150 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
151 adjacent to the end of the range, and the range contained characters with
152 more than one other case, caused incorrect behaviour when compiled in UTF
153 mode. In that example, the range a-j was left out of the class.
155 26. Fix JIT compilation of conditional blocks, which assertion
156 is converted to (*FAIL). E.g: /(?(?!))/.
158 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
159 discovered by the LLVM fuzzer.
161 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
162 when this assertion was used as a condition, for example (?(?!)a|b). In
163 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
164 error about an unsupported item.
166 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
167 possessification code could take exponential time to complete. A recursion
168 depth limit of 1000 has been imposed to limit the resources used by this
169 optimization.
171 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
172 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
173 because \S ensures they are all in the class. The code for doing this was
174 interacting badly with the code for computing the amount of space needed to
175 compile the pattern, leading to a buffer overflow. This bug was discovered
176 by the LLVM fuzzer.
178 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
179 other kinds of group caused stack overflow at compile time. This bug was
180 discovered by the LLVM fuzzer.
182 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
183 between a subroutine call and its quantifier was incorrectly compiled,
184 leading to buffer overflow or other errors. This bug was discovered by the
185 LLVM fuzzer.
187 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
188 assertion after (?(. The code was failing to check the character after
189 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
190 was discovered by the LLVM fuzzer.
192 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
193 a fixed maximum following a group that contains a subroutine reference was
194 incorrectly compiled and could trigger buffer overflow. This bug was
195 discovered by the LLVM fuzzer.
197 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
198 caused a stack overflow instead of the diagnosis of a non-fixed length
199 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
201 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
202 (e.g. /(?<=\Ka)/) could make pcregrep loop.
204 37. There was a similar problem to 36 in pcretest for global matches.
206 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
207 and a subsequent item in the pattern caused a non-match, backtracking over
208 the repeated \X did not stop, but carried on past the start of the subject,
209 causing reference to random memory and/or a segfault. There were also some
210 other cases where backtracking after \C could crash. This set of bugs was
211 discovered by the LLVM fuzzer.
213 39. The function for finding the minimum length of a matching string could take
214 a very long time if mutual recursion was present many times in a pattern,
215 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
216 method has been implemented. This infelicity was discovered by the LLVM
217 fuzzer.
219 40. Static linking against the PCRE library using the pkg-config module was
220 failing on missing pthread symbols.
223 Version 8.36 26-September-2014
224 ------------------------------
226 1. Got rid of some compiler warnings in the C++ modules that were shown up by
227 -Wmissing-field-initializers and -Wunused-parameter.
229 2. The tests for quantifiers being too big (greater than 65535) were being
230 applied after reading the number, and stupidly assuming that integer
231 overflow would give a negative number. The tests are now applied as the
232 numbers are read.
234 3. Tidy code in pcre_exec.c where two branches that used to be different are
235 now the same.
237 4. The JIT compiler did not generate match limit checks for certain
238 bracketed expressions with quantifiers. This may lead to exponential
239 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
240 issue should be resolved now.
242 5. Fixed an issue, which occures when nested alternatives are optimized
243 with table jumps.
245 6. Inserted two casts and changed some ints to size_t in the light of some
246 reported 64-bit compiler warnings (Bugzilla 1477).
248 7. Fixed a bug concerned with zero-minimum possessive groups that could match
249 an empty string, which sometimes were behaving incorrectly in the
250 interpreter (though correctly in the JIT matcher). This pcretest input is
251 an example:
253 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
256 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
257 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
258 for an empty string was breaking the inner loop and carrying on at a lower
259 level, when possessive repeated groups should always return to a higher
260 level as they have no backtrack points in them. The empty string test now
261 occurs at the outer level.
263 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
264 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
266 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
267 doesn't).
269 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
270 to hold a pattern to be incorrectly computed (too small) when there were
271 named back references to duplicated names. This could cause "internal
272 error: code overflow" or "double free or corruption" or other memory
273 handling errors.
275 11. When named subpatterns had the same prefixes, back references could be
276 confused. For example, in this pattern:
278 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
280 the reference to 'Name' was incorrectly treated as a reference to a
281 duplicate name.
283 12. A pattern such as /^s?c/mi8 where the optional character has more than
284 one "other case" was incorrectly compiled such that it would only try to
285 match starting at "c".
287 13. When a pattern starting with \s was studied, VT was not included in the
288 list of possible starting characters; this should have been part of the
289 8.34/18 patch.
291 14. If a character class started [\Qx]... where x is any character, the class
292 was incorrectly terminated at the ].
294 15. If a pattern that started with a caseless match for a character with more
295 than one "other case" was studied, PCRE did not set up the starting code
296 unit bit map for the list of possible characters. Now it does. This is an
297 optimization improvement, not a bug fix.
299 16. The Unicode data tables have been updated to Unicode 7.0.0.
301 17. Fixed a number of memory leaks in pcregrep.
303 18. Avoid a compiler warning (from some compilers) for a function call with
304 a cast that removes "const" from an lvalue by using an intermediate
305 variable (to which the compiler does not object).
307 19. Incorrect code was compiled if a group that contained an internal recursive
308 back reference was optional (had quantifier with a minimum of zero). This
309 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
310 caused segmentation faults because of stack overflows at compile time.
312 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
313 group that is quantified with an indefinite repeat, caused a compile-time
314 loop which used up all the system stack and provoked a segmentation fault.
315 This was not the same bug as 19 above.
317 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
318 Patch by Mike Frysinger.
321 Version 8.35 04-April-2014
322 --------------------------
324 1. A new flag is set, when property checks are present in an XCLASS.
325 When this flag is not set, PCRE can perform certain optimizations
326 such as studying these XCLASS-es.
328 2. The auto-possessification of character sets were improved: a normal
329 and an extended character set can be compared now. Furthermore
330 the JIT compiler optimizes more character set checks.
332 3. Got rid of some compiler warnings for potentially uninitialized variables
333 that show up only when compiled with -O2.
335 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
336 of a match later then the end of the match. The pcretest program was not
337 handling the case sensibly - it was outputting from the start to the next
338 binary zero. It now reports this situation in a message, and outputs the
339 text from the end to the start.
341 5. Fast forward search is improved in JIT. Instead of the first three
342 characters, any three characters with fixed position can be searched.
343 Search order: first, last, middle.
345 6. Improve character range checks in JIT. Characters are read by an inprecise
346 function now, which returns with an unknown value if the character code is
347 above a certain threshold (e.g: 256). The only limitation is that the value
348 must be bigger than the threshold as well. This function is useful when
349 the characters above the threshold are handled in the same way.
351 7. The macros whose names start with RAWUCHAR are placeholders for a future
352 mode in which only the bottom 21 bits of 32-bit data items are used. To
353 make this more memorable for those maintaining the code, the names have
354 been changed to start with UCHAR21, and an extensive comment has been added
355 to their definition.
357 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
358 to the export list in Makefile.am (they were accidentally omitted from the
359 8.34 tarball).
361 9. The informational output from pcretest used the phrase "starting byte set"
362 which is inappropriate for the 16-bit and 32-bit libraries. As the output
363 for "first char" and "need char" really means "non-UTF-char", I've changed
364 "byte" to "char", and slightly reworded the output. The documentation about
365 these values has also been (I hope) clarified.
367 10. Another JIT related optimization: use table jumps for selecting the correct
368 backtracking path, when more than four alternatives are present inside a
369 bracket.
371 11. Empty match is not possible, when the minimum length is greater than zero,
372 and there is no \K in the pattern. JIT should avoid empty match checks in
373 such cases.
375 12. In a caseless character class with UCP support, when a character with more
376 than one alternative case was not the first character of a range, not all
377 the alternative cases were added to the class. For example, s and \x{17f}
378 are both alternative cases for S: the class [RST] was handled correctly,
379 but [R-T] was not.
381 13. The configure.ac file always checked for pthread support when JIT was
382 enabled. This is not used in Windows, so I have put this test inside a
383 check for the presence of windows.h (which was already tested for).
385 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
386 The algorithm provides a way to skip certain starting offsets, and usually
387 faster than linear prefix searches.
389 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
390 as for 'fr_FR' and 'french'. For some reason, however, it then used the
391 Windows-specific input and output files, which have 'french' screwed in.
392 So this could never have worked. One of the problems with locales is that
393 they aren't always the same. I have now updated RunTest so that it checks
394 the output of the locale test (test 3) against three different output
395 files, and it allows the test to pass if any one of them matches. With luck
396 this should make the test pass on some versions of Solaris where it was
397 failing. Because of the uncertainty, the script did not used to stop if
398 test 3 failed; it now does. If further versions of a French locale ever
399 come to light, they can now easily be added.
401 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
402 there was a message during ./configure, but it did not stop. This now
403 provokes an error. The invalid example in README has been corrected.
404 If a value less than the minimum is given, the minimum value has always
405 been used, but now a warning is given.
407 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
408 was a bug in the test system, which is now fixed. Also, the list of various
409 configurations that are tested for each release did not have one with both
410 16/32 bits and --enable-bar-anycrlf. It now does.
412 18. pcretest was missing "-C bsr" for displaying the \R default setting.
414 19. Little endian PowerPC systems are supported now by the JIT compiler.
416 20. The fast forward newline mechanism could enter to an infinite loop on
417 certain invalid UTF-8 input. Although we don't support these cases
418 this issue can be fixed by a performance optimization.
420 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
421 not take account if existing stack usage. There is now a new global
422 variable called pcre_stack_guard that can be set to point to an external
423 function to check stack availability. It is called at the start of
424 processing every parenthesized group.
426 22. A typo in the code meant that in ungreedy mode the max/min qualifier
427 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
428 not match "ab".
430 23. When UTF was disabled, the JIT program reported some incorrect compile
431 errors. These messages are silenced now.
433 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
434 compiler.
436 25. Change all the temporary files used in RunGrepTest to be different to those
437 used by RunTest so that the tests can be run simultaneously, for example by
438 "make -j check".
441 Version 8.34 15-December-2013
442 -----------------------------
444 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
445 executable memory. Patch inspired by Carsten Klein.
447 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
448 this macro is never tested and has no effect, because the work to support
449 coverage involves only compiling and linking options and special targets in
450 the Makefile. The comment in config.h implied that defining the macro would
451 enable coverage support, which is totally false. There was also support for
452 setting this macro in the CMake files (my fault, I just copied it from
453 configure). SUPPORT_GCOV has now been removed.
455 3. Make a small performance improvement in strlen16() and strlen32() in
456 pcretest.
458 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
459 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
460 these cases). There was also one in pcretest.c.
462 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
464 6. In UTF mode, the code for checking whether a group could match an empty
465 string (which is used for indefinitely repeated groups to allow for
466 breaking an infinite loop) was broken when the group contained a repeated
467 negated single-character class with a character that occupied more than one
468 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
469 UTF-8 mode). The effect was undefined: the group might or might not be
470 deemed as matching an empty string, or the program might have crashed.
472 7. The code for checking whether a group could match an empty string was not
473 recognizing that \h, \H, \v, \V, and \R must match a character.
475 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
476 an empty string. If it can, pcretest shows this in its information output.
478 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
479 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
480 matched by pcre_exec() without using JIT:
482 (a) If the rest of the pattern did not match after a maximal run of
483 grapheme clusters, the code for backing up to try with fewer of them
484 did not always back up over a full grapheme when characters that do not
485 have the modifier quality were involved, e.g. Hangul syllables.
487 (b) If the match point in a subject started with modifier character, and
488 there was no match, the code could incorrectly back up beyond the match
489 point, and potentially beyond the first character in the subject,
490 leading to a segfault or an incorrect match result.
492 10. A conditional group with an assertion condition could lead to PCRE
493 recording an incorrect first data item for a match if no other first data
494 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
495 first data item, and therefore matched "ca" after "c" instead of at the
496 start.
498 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
499 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
501 12. The source of pcregrep now includes z/OS-specific code so that it can be
502 compiled for z/OS as part of the special z/OS distribution.
504 13. Added the -T and -TM options to pcretest.
506 14. The code in pcre_compile.c for creating the table of named capturing groups
507 has been refactored. Instead of creating the table dynamically during the
508 actual compiling pass, the information is remembered during the pre-compile
509 pass (on the stack unless there are more than 20 named groups, in which
510 case malloc() is used) and the whole table is created before the actual
511 compile happens. This has simplified the code (it is now nearly 150 lines
512 shorter) and prepared the way for better handling of references to groups
513 with duplicate names.
515 15. A back reference to a named subpattern when there is more than one of the
516 same name now checks them in the order in which they appear in the pattern.
517 The first one that is set is used for the reference. Previously only the
518 first one was inspected. This change makes PCRE more compatible with Perl.
520 16. Unicode character properties were updated from Unicode 6.3.0.
522 17. The compile-time code for auto-possessification has been refactored, based
523 on a patch by Zoltan Herczeg. It now happens after instead of during
524 compilation. The code is cleaner, and more cases are handled. The option
525 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
526 options in pcretest are provided to set it. It can also be set by
527 (*NO_AUTO_POSSESS) at the start of a pattern.
529 18. The character VT has been added to the default ("C" locale) set of
530 characters that match \s and are generally treated as white space,
531 following this same change in Perl 5.18. There is now no difference between
532 "Perl space" and "POSIX space". Whether VT is treated as white space in
533 other locales depends on the locale.
535 19. The code for checking named groups as conditions, either for being set or
536 for being recursed, has been refactored (this is related to 14 and 15
537 above). Processing unduplicated named groups should now be as fast at
538 numerical groups, and processing duplicated groups should be faster than
539 before.
541 20. Two patches to the CMake build system, by Alexander Barkov:
543 (1) Replace the "source" command by "." in CMakeLists.txt because
544 "source" is a bash-ism.
546 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
547 without these the CMake build does not work on Solaris.
549 21. Perl has changed its handling of \8 and \9. If there is no previously
550 encountered capturing group of those numbers, they are treated as the
551 literal characters 8 and 9 instead of a binary zero followed by the
552 literals. PCRE now does the same.
554 22. Following Perl, added \o{} to specify codepoints in octal, making it
555 possible to specify values greater than 0777 and also making them
556 unambiguous.
558 23. Perl now gives an error for missing closing braces after \x{... instead of
559 treating the string as literal. PCRE now does the same.
561 24. RunTest used to grumble if an inappropriate test was selected explicitly,
562 but just skip it when running all tests. This make it awkward to run ranges
563 of tests when one of them was inappropriate. Now it just skips any
564 inappropriate tests, as it always did when running all tests.
566 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
567 character types such as \d or \w, too many callouts were inserted, and the
568 data that they returned was rubbish.
570 26. In UCP mode, \s was not matching two of the characters that Perl matches,
571 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
572 were matched by \h. The code has now been refactored so that the lists of
573 the horizontal and vertical whitespace characters used for \h and \v (which
574 are defined only in one place) are now also used for \s.
576 27. Add JIT support for the 64 bit TileGX architecture.
577 Patch by Jiong Wang (Tilera Corporation).
579 28. Possessive quantifiers for classes (both explicit and automatically
580 generated) now use special opcodes instead of wrapping in ONCE brackets.
582 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
583 (because it's meaningless), this was not happening when PCRE_CASELESS was
584 set. Not wrong, but inefficient.
586 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
587 properties for \w, \d, etc) is present in a test regex. Otherwise if the
588 test contains no characters greater than 255, Perl doesn't realise it
589 should be using Unicode semantics.
591 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
592 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
593 does in Unicode mode.
595 32. Added the "forbid" facility to pcretest so that putting tests into the
596 wrong test files can sometimes be quickly detected.
598 33. There is now a limit (default 250) on the depth of nesting of parentheses.
599 This limit is imposed to control the amount of system stack used at compile
600 time. It can be changed at build time by --with-parens-nest-limit=xxx or
601 the equivalent in CMake.
603 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
604 errors. Perl warns for these when in warning mode, but PCRE has no facility
605 for giving warnings.
607 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
608 However, this was not working for (?!) because it is optimized to (*FAIL),
609 for which PCRE does not allow quantifiers. The optimization is now disabled
610 when a quantifier follows (?!). I can't see any use for this, but it makes
611 things uniform.
613 36. Perl no longer allows group names to start with digits, so I have made this
614 change also in PCRE. It simplifies the code a bit.
616 37. In extended mode, Perl ignores spaces before a + that indicates a
617 possessive quantifier. PCRE allowed a space before the quantifier, but not
618 before the possessive +. It now does.
620 38. The use of \K (reset reported match start) within a repeated possessive
621 group such as (a\Kb)*+ was not working.
623 40. Document that the same character tables must be used at compile time and
624 run time, and that the facility to pass tables to pcre_exec() and
625 pcre_dfa_exec() is for use only with saved/restored patterns.
627 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
628 features for Builds with MSVC:
630 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
631 builds). The .rc files can be used to set FileDescription and many other
632 attributes.
634 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
635 This allows higher-level build scripts which want .pdb files to avoid
636 hard-coding the exact files needed."
638 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
639 mean "start of word" and "end of word", respectively, as a transition aid.
641 43. A minimizing repeat of a class containing codepoints greater than 255 in
642 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
643 compiled to use the heap for recursion.
645 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
646 is configured.
649 Version 8.33 28-May-2013
650 ------------------------
652 1. Added 'U' to some constants that are compared to unsigned integers, to
653 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
654 variables that are added to signed variables, to ensure the result is
655 signed and can be negated.
657 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
658 CMake config files.
660 3. Revise the creation of config.h.generic so that all boolean macros are
661 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
662 overriding via -D on the command line possible.
664 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
665 to unsigned int is reported to make a quite noticeable speed difference in
666 a specific Windows environment. Testing on Linux did also appear to show
667 some benefit (and it is clearly not harmful). Also fixed the definition of
668 Xop which should be unsigned.
670 5. Related to (4), changing the definition of the intermediate variable cc
671 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
672 improvements.
674 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
675 unnecessary spaces.
677 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
678 and later.
680 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
682 9. Optimizing fast_forward_start_bits in JIT.
684 10. Adding support for callouts in JIT, and fixing some issues revealed
685 during this work. Namely:
687 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
689 (b) Minimum length was not checked before the matching is started.
691 11. The value of capture_last that is passed to callouts was incorrect in some
692 cases when there was a capture on one path that was subsequently abandoned
693 after a backtrack. Also, the capture_last value is now reset after a
694 recursion, since all captures are also reset in this case.
696 12. The interpreter no longer returns the "too many substrings" error in the
697 case when an overflowing capture is in a branch that is subsequently
698 abandoned after a backtrack.
700 13. In the pathological case when an offset vector of size 2 is used, pcretest
701 now prints out the matched string after a yield of 0 or 1.
703 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
704 Only supported by the JIT compiler at the moment.
706 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
708 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
709 offset of the starting point of the matching process, provided the offsets
710 vector is large enough.
712 17. The \A escape now records a lookbehind value of 1, though its execution
713 does not actually inspect the previous character. This is to ensure that,
714 in partial multi-segment matching, at least one character from the old
715 segment is retained when a new segment is processed. Otherwise, if there
716 are no lookbehinds in the pattern, \A might match incorrectly at the start
717 of a new segment.
719 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
721 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
722 gives some modest performance improvement in 8-bit mode.
724 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
725 be expressed in certain programming languages using Universal Character
726 Names.
728 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
729 which points out that "non characters" are not "characters that may not
730 appear in Unicode strings" but rather "characters that are reserved for
731 internal use and have only local meaning".
733 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
734 there was a conditional group that depended on an assertion, if the
735 assertion was false, the callout that immediately followed the alternation
736 in the condition was skipped when pcre_exec() was used for matching.
738 23. Allow an explicit callout to be inserted before an assertion that is the
739 condition for a conditional group, for compatibility with automatic
740 callouts, which always insert a callout at this point.
742 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
743 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
745 25. (*PRUNE) is now supported by the JIT compiler.
747 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
749 27. Fix the case where there are two or more SKIPs with arguments that may be
750 ignored.
752 28. (*SKIP) is now supported by the JIT compiler.
754 29. (*THEN) is now supported by the JIT compiler.
756 30. Update RunTest with additional test selector options.
758 31. The way PCRE handles backtracking verbs has been changed in two ways.
760 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
761 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
762 backtracking. In some cases this makes it more Perl-compatible, but Perl's
763 rather obscure rules do not always do the same thing.
765 (2) Previously, backtracking verbs were confined within assertions. This is
766 no longer the case for positive assertions, except for (*ACCEPT). Again,
767 this sometimes improves Perl compatibility, and sometimes does not.
769 32. A number of tests that were in test 2 because Perl did things differently
770 have been moved to test 1, because either Perl or PCRE has changed, and
771 these tests are now compatible.
773 32. Backtracking control verbs are now handled in the same way in JIT and
774 interpreter.
776 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
777 contained a forward subroutine reference caused a compile error.
779 34. Auto-detect and optimize limited repetitions in JIT.
781 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
782 blocking (*UTF) etc.
784 36. In the interpreter, maximizing pattern repetitions for characters and
785 character types now use tail recursion, which reduces stack usage.
787 37. The value of the max lookbehind was not correctly preserved if a compiled
788 and saved regex was reloaded on a host of different endianness.
790 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
791 of the compiled pattern block, expand the flags field from 16 to 32 bits
792 because it was almost full.
794 39. Try madvise first before posix_madvise.
796 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
797 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
798 original change locked out all matches of empty strings. This has been
799 changed so that one match of an empty string per line is recognized.
800 Subsequent searches on the same line (for colouring or for --only-matching,
801 for example) do not recognize empty strings.
803 41. Applied a user patch to fix a number of spelling mistakes in comments.
805 42. Data lines longer than 65536 caused pcretest to crash.
807 43. Clarified the data type for length and startoffset arguments for pcre_exec
808 and pcre_dfa_exec in the function-specific man pages, where they were
809 explicitly stated to be in bytes, never having been updated. I also added
810 some clarification to the pcreapi man page.
812 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
813 a segmentation fault.
816 Version 8.32 30-November-2012
817 -----------------------------
819 1. Improved JIT compiler optimizations for first character search and single
820 character iterators.
822 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
823 Patch by Daniel Richard G.
825 3. Single character iterator optimizations in the JIT compiler.
827 4. Improved JIT compiler optimizations for character ranges.
829 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
830 Reported by Giuseppe D'Angelo.
832 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
833 of a line, was being set incorrectly in cases where .* appeared inside
834 atomic brackets at the start of a pattern, or where there was a subsequent
835 *PRUNE or *SKIP.
837 7. Improved instruction cache flush for POWER/PowerPC.
838 Patch by Daniel Richard G.
840 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
841 grep:
843 (a) There is now no limit to the number of patterns to be matched.
845 (b) An error is given if a pattern is too long.
847 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
848 are now supported.
850 (d) --exclude-from and --include-from (multiple use) have been added.
852 (e) Exclusions and inclusions now apply to all files and directories, not
853 just to those obtained from scanning a directory recursively.
855 (f) Multiple uses of -f and --file-list are now supported.
857 (g) In a Windows environment, the default for -d has been changed from
858 "read" (the GNU grep default) to "skip", because otherwise the presence
859 of a directory in the file list provokes an error.
861 (h) The documentation has been revised and clarified in places.
863 9. Improve the matching speed of capturing brackets.
865 10. Changed the meaning of \X so that it now matches a Unicode extended
866 grapheme cluster.
868 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
869 out POSIX threads when JIT support is configured.
871 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
873 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
874 structure could go wrong in environments where size_t is not the same size
875 as int.
877 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
878 set.
880 15. The EBCDIC support had decayed; later updates to the code had included
881 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
882 general tidy up of EBCDIC-related issues, and the documentation was also
883 not quite right. There is now a test that can be run on ASCII systems to
884 check some of the EBCDIC-related things (but is it not a full test).
886 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
887 in a small tidy to the code.
889 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
891 18. If the --only-matching (-o) option in pcregrep is specified multiple
892 times, each one causes appropriate output. For example, -o1 -o2 outputs the
893 substrings matched by the 1st and 2nd capturing parentheses. A separating
894 string can be specified by --om-separator (default empty).
896 19. Improving the first n character searches.
898 20. Turn case lists for horizontal and vertical white space into macros so that
899 they are defined only once.
901 21. This set of changes together give more compatible Unicode case-folding
902 behaviour for characters that have more than one other case when UCP
903 support is available.
905 (a) The Unicode property table now has offsets into a new table of sets of
906 three or more characters that are case-equivalent. The MultiStage2.py
907 script that generates these tables (the pcre_ucd.c file) now scans
908 CaseFolding.txt instead of UnicodeData.txt for character case
909 information.
911 (b) The code for adding characters or ranges of characters to a character
912 class has been abstracted into a generalized function that also handles
913 case-independence. In UTF-mode with UCP support, this uses the new data
914 to handle characters with more than one other case.
916 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
917 whose other case is greater than 256 are now correctly matched
918 caselessly. Previously, the high codepoint matched the low one, but not
919 vice versa.
921 (d) The processing of \h, \H, \v, and \ in character classes now makes use
922 of the new class addition function, using character lists defined as
923 macros alongside the case definitions of 20 above.
925 (e) Caseless back references now work with characters that have more than
926 one other case.
928 (f) General caseless matching of characters with more than one other case
929 is supported.
931 22. Unicode character properties were updated from Unicode 6.2.0
933 23. Improved CMake support under Windows. Patch by Daniel Richard G.
935 24. Add support for 32-bit character strings, and UTF-32
937 25. Major JIT compiler update (code refactoring and bugfixing).
938 Experimental Sparc 32 support is added.
940 26. Applied a modified version of Daniel Richard G's patch to create
941 pcre.h.generic and config.h.generic by "make" instead of in the
942 PrepareRelease script.
944 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
945 pcre_compile.c when checking for a zero character.
947 28. Introducing a native interface for JIT. Through this interface, the compiled
948 machine code can be directly executed. The purpose of this interface is to
949 provide fast pattern matching, so several sanity checks are not performed.
950 However, feature tests are still performed. The new interface provides
951 1.4x speedup compared to the old one.
953 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
954 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
955 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
957 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
958 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
959 the "old" RFC 2279). Instead, it ended up passing a negative length to
960 pcre_exec().
962 31. Add support for GCC's visibility feature to hide internal functions.
964 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
965 "unknown -C option" after outputting 0 or 1.
967 33. There is now support for generating a code coverage report for the test
968 suite in environments where gcc is the compiler and lcov is installed. This
969 is mainly for the benefit of the developers.
971 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
972 unaddressable using valgrind annotations, allowing valgrind to detect
973 invalid memory accesses. This is mainly for the benefit of the developers.
975 25. (*UTF) can now be used to start a pattern in any of the three libraries.
977 26. Give configure error if --enable-cpp but no C++ compiler found.
980 Version 8.31 06-July-2012
981 -------------------------
983 1. Fixing a wrong JIT test case and some compiler warnings.
985 2. Removed a bashism from the RunTest script.
987 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
988 to unsigned type, result still unsigned" that was given by an MS compiler
989 on encountering the code "-sizeof(xxx)".
991 4. Partial matching support is added to the JIT compiler.
993 5. Fixed several bugs concerned with partial matching of items that consist
994 of more than one character:
996 (a) /^(..)\1/ did not partially match "aba" because checking references was
997 done on an "all or nothing" basis. This also applied to repeated
998 references.
1000 (b) \R did not give a hard partial match if \r was found at the end of the
1001 subject.
1003 (c) \X did not give a hard partial match after matching one or more
1004 characters at the end of the subject.
1006 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1007 a partial match for the string "\r".
1009 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1010 a partial match for a CR character at the end of the subject string.
1012 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1013 running pcretest, the text "(JIT)" added to the output whenever JIT is
1014 actually used to run the match.
1016 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1017 or /S+[+] with a digit between 1 and 7.
1019 8. OP_NOT now supports any UTF character not just single-byte ones.
1021 9. (*MARK) control verb is now supported by the JIT compiler.
1023 10. The command "./RunTest list" lists the available tests without actually
1024 running any of them. (Because I keep forgetting what they all are.)
1028 12. Applied a (slightly modified) user-supplied patch that improves performance
1029 when the heap is used for recursion (compiled with --disable-stack-for-
1030 recursion). Instead of malloc and free for each heap frame each time a
1031 logical recursion happens, frames are retained on a chain and re-used where
1032 possible. This sometimes gives as much as 30% improvement.
1034 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1035 call.
1037 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1039 15. It is now possible to link pcretest with libedit as an alternative to
1040 libreadline.
1042 16. (*COMMIT) control verb is now supported by the JIT compiler.
1044 17. The Unicode data tables have been updated to Unicode 6.1.0.
1046 18. Added --file-list option to pcregrep.
1048 19. Added binary file support to pcregrep, including the -a, --binary-files,
1049 -I, and --text options.
1051 20. The madvise function is renamed for posix_madvise for QNX compatibility
1052 reasons. Fixed by Giuseppe D'Angelo.
1054 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1056 22. Changed the option for creating symbolic links for 16-bit man pages from
1057 -s to -sf so that re-installing does not cause issues.
1059 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1061 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1062 matches in certain environments (the workspace was not being correctly
1063 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1064 some of the workspace data at the beginning of a restart.
1066 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1067 was not doing so when it should - probably a typo introduced by SVN 528
1068 (change 8.10/14).
1070 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1071 \w+ when the character tables indicated that \x{c4} was a word character.
1072 There were several related cases, all because the tests for doing a table
1073 lookup were testing for characters less than 127 instead of 255.
1075 27. If a pattern contains capturing parentheses that are not used in a match,
1076 their slots in the ovector are set to -1. For those that are higher than
1077 any matched groups, this happens at the end of processing. In the case when
1078 there were back references that the ovector was too small to contain
1079 (causing temporary malloc'd memory to be used during matching), and the
1080 highest capturing number was not used, memory off the end of the ovector
1081 was incorrectly being set to -1. (It was using the size of the temporary
1082 memory instead of the true size.)
1084 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1085 ovector size, it uses memory at the end of the block that it has got.
1087 29. Check for an overlong MARK name and give an error at compile time. The
1088 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1090 30. JIT compiler update.
1092 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1093 Rill for the patch.
1095 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1097 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1099 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1100 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1101 was enabled.
1103 35. Improve JIT code generation for greedy plus quantifier.
1105 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1106 1 to "aa" instead of to an empty string. The bug affected repeated groups
1107 that could potentially match an empty string.
1109 37. Optimizing single character iterators in JIT.
1111 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1112 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1113 codepoints that are too big for the mode are faulted, and in a UTF mode,
1114 disallowed codepoints are also faulted.
1116 39. If PCRE was compiled with UTF support, in three places in the DFA
1117 matcher there was code that should only have been obeyed in UTF mode, but
1118 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1119 processing when bytes with values greater than 127 were present. In 16-bit
1120 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1121 both cases the values are those that cannot be the first data item in a UTF
1122 character. The three items that might have provoked this were recursions,
1123 possessively repeated groups, and atomic groups.
1125 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1126 and pcregrep, because some OS require shared objects to be explicitly
1127 passed to ld, causing the link step to fail if they are not.
1129 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1130 mode, patterns that started with \h* or \R* might be incorrectly matched.
1133 Version 8.30 04-February-2012
1134 -----------------------------
1136 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1137 name is defined in ctype.h.
1139 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1140 only in quite long subpatterns.
1142 3. Removed the function pcre_info(), which has been obsolete and deprecated
1143 since it was replaced by pcre_fullinfo() in February 2000.
1145 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1146 match a (*MARK), and the match failed at the start of the subject, a
1147 reference to memory before the start of the subject could occur. This bug
1148 was introduced by fix 17 of release 8.21.
1150 5. A reference to an unset group with zero minimum repetition was giving
1151 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1152 /(another)?(\1?)test/ matched against "hello world test". This bug was
1153 introduced in release 8.13.
1155 6. Add support for 16-bit character strings (a large amount of work involving
1156 many changes and refactorings).
1158 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1159 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1160 from a file.
1162 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1163 rounding is not applied in this particular case).
1165 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1166 if they appear, or are escaped, in patterns.
1168 10. Get rid of a number of -Wunused-but-set-variable warnings.
1170 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1171 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1172 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1173 also returns the mark "x". This bug applied to capturing parentheses,
1174 non-capturing parentheses, and atomic parentheses. It also applied to some
1175 assertions.
1177 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1178 information out of configure.ac instead of relying on pcre.h.generic, which
1179 is not stored in the repository.
1181 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1182 -lreadline.
1184 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1186 15. Applied Graycode's patch to put the top-level frame on the stack rather
1187 than the heap when not using the stack for recursion. This gives a
1188 performance improvement in many cases when recursion is not deep.
1190 16. Experimental code added to "pcretest -C" to output the stack frame size.
1193 Version 8.21 12-Dec-2011
1194 ------------------------
1196 1. Updating the JIT compiler.
1198 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1199 are added as well.
1201 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1202 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1203 calling _pcre_jit_exec. Some extra comments are added.
1205 4. (*MARK) settings inside atomic groups that do not contain any capturing
1206 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1207 was introduced by change 18 for 8.20.
1209 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1210 ECMA-262 standard.
1212 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1213 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1214 This bug was probably introduced by change 9 of 8.13.
1216 7. While fixing 6 above, I noticed that a number of other items were being
1217 incorrectly rejected as "not fixed length". This arose partly because newer
1218 opcodes had not been added to the fixed-length checking code. I have (a)
1219 corrected the bug and added tests for these items, and (b) arranged for an
1220 error to occur if an unknown opcode is encountered while checking for fixed
1221 length instead of just assuming "not fixed length". The items that were
1222 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1223 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1224 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1226 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1227 being incorrectly compiled and would have given unpredicatble results.
1229 9. A possessively repeated subpattern with minimum repeat count greater than
1230 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1231 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1232 the first (A) could occur when it should not.
1234 10. Add a cast and remove a redundant test from the code.
1236 11. JIT should use pcre_malloc/pcre_free for allocation.
1238 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1239 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1240 is anything other than /usr, -L is still shown).
1242 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1244 14. Perl does not support \N without a following name in a [] class; PCRE now
1245 also gives an error.
1247 15. If a forward reference was repeated with an upper limit of around 2000,
1248 it caused the error "internal error: overran compiling workspace". The
1249 maximum number of forward references (including repeats) was limited by the
1250 internal workspace, and dependent on the LINK_SIZE. The code has been
1251 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1252 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1253 of around 200,000 forward references. While doing this, I also speeded up
1254 the filling in of repeated forward references.
1256 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1257 incorrectly expecting the subject to contain another "a" after the start.
1259 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1260 in the match, the SKIP should be ignored. This was not happening; instead
1261 the SKIP was being treated as NOMATCH. For patterns such as
1262 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1263 tested.
1265 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1266 now much more compatible with Perl, in particular in cases where the result
1267 is a non-match for a non-anchored pattern. For example, if
1268 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1269 "m", where previously it did not return a name. A side effect of this
1270 change is that for partial matches, the last encountered mark name is
1271 returned, as for non matches. A number of tests that were previously not
1272 Perl-compatible have been moved into the Perl-compatible test files. The
1273 refactoring has had the pleasing side effect of removing one argument from
1274 the match() function, thus reducing its stack requirements.
1276 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1277 subsequent uses of /S (without +) incorrectly behaved like /S+.
1279 21. Retrieve executable code size support for the JIT compiler and fixing
1280 some warnings.
1282 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1283 not work when the shorter character appeared right at the end of the
1284 subject string.
1286 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1287 systems.
1289 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1290 output it when the /M option is used in pcretest.
1292 25. The CheckMan script was not being included in the distribution. Also, added
1293 an explicit "perl" to run Perl scripts from the PrepareRelease script
1294 because this is reportedly needed in Windows.
1296 26. If study data was being save in a file and studying had not found a set of
1297 "starts with" bytes for the pattern, the data written to the file (though
1298 never used) was taken from uninitialized memory and so caused valgrind to
1299 complain.
1301 27. Updated RunTest.bat as provided by Sheri Pierce.
1303 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1305 29. Computation of memory usage for the table of capturing group names was
1306 giving an unnecessarily large value.
1309 Version 8.20 21-Oct-2011
1310 ------------------------
1312 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1313 a POSIX class. After further experiments with Perl, which convinced me that
1314 Perl has bugs and confusions, a closing square bracket is no longer allowed
1315 in a POSIX name. This bug also affected patterns with classes that started
1316 with full stops.
1318 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1319 captured substring, but while checking the failing first alternative,
1320 substring 1 is temporarily captured. If the output vector supplied to
1321 pcre_exec() was not big enough for this capture, the yield of the function
1322 was still zero ("insufficient space for captured substrings"). This cannot
1323 be totally fixed without adding another stack variable, which seems a lot
1324 of expense for a edge case. However, I have improved the situation in cases
1325 such as /(a)(b)x|abc/ matched against "abc", where the return code
1326 indicates that fewer than the maximum number of slots in the ovector have
1327 been set.
1329 3. Related to (2) above: when there are more back references in a pattern than
1330 slots in the output vector, pcre_exec() uses temporary memory during
1331 matching, and copies in the captures as far as possible afterwards. It was
1332 using the entire output vector, but this conflicts with the specification
1333 that only 2/3 is used for passing back captured substrings. Now it uses
1334 only the first 2/3, for compatibility. This is, of course, another edge
1335 case.
1337 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1338 main code base, and can be used by building with --enable-jit. When this is
1339 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1340 runtime --no-jit option is given.
1342 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1343 ovector, the return from the function was zero, implying that there were
1344 other matches that did not fit. The correct "exactly full" value is now
1345 returned.
1347 6. If a subpattern that was called recursively or as a subroutine contained
1348 (*PRUNE) or any other control that caused it to give a non-standard return,
1349 invalid errors such as "Error -26 (nested recursion at the same subject
1350 position)" or even infinite loops could occur.
1352 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1353 computing the minimum length on reaching *ACCEPT, and so ended up with the
1354 wrong value of 1 rather than 0. Further investigation indicates that
1355 computing a minimum subject length in the presence of *ACCEPT is difficult
1356 (think back references, subroutine calls), and so I have changed the code
1357 so that no minimum is registered for a pattern that contains *ACCEPT.
1359 8. If (*THEN) was present in the first (true) branch of a conditional group,
1360 it was not handled as intended. [But see 16 below.]
1362 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1363 Sheri Pierce.
1365 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1366 the first byte in a match must be "a".
1368 11. Change 17 for 8.13 increased the recursion depth for patterns like
1369 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1370 pattern contains any instances of (*THEN). If it does not, the old
1371 optimizations are restored. It would be nice to do this on a per-group
1372 basis, but at the moment that is not feasible.
1374 12. In some environments, the output of pcretest -C is CRLF terminated. This
1375 broke RunTest's code that checks for the link size. A single white space
1376 character after the value is now allowed for.
1378 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1379 For "fr", it uses the Windows-specific input and output files.
1381 14. If (*THEN) appeared in a group that was called recursively or as a
1382 subroutine, it did not work as intended. [But see next item.]
1384 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1385 pattern fragments (but not containing any | characters). If A and B are
1386 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1387 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1388 D. In other words, Perl considers parentheses that do not contain any |
1389 characters to be part of a surrounding alternative, whereas PCRE was
1390 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1391 differently. PCRE now behaves in the same way as Perl, except in the case
1392 of subroutine/recursion calls such as (?1) which have in any case always
1393 been different (but PCRE had them first :-).
1395 16. Related to 15 above: Perl does not treat the | in a conditional group as
1396 creating alternatives. Such a group is treated in the same way as an
1397 ordinary group without any | characters when processing (*THEN). PCRE has
1398 been changed to match Perl's behaviour.
1400 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1401 RunGrepTest script failed.
1403 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1404 inevitable for groups that contain captures, but it can lead to a lot of
1405 stack use in large patterns. The old behaviour has been restored for atomic
1406 groups that do not contain any capturing parentheses.
1408 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1409 suppress the check for a minimum subject length at run time. (If it was
1410 given to pcre_exec() or pcre_dfa_exec() it did work.)
1412 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1413 fail to work when decoding hex characters in data strings in EBCDIC
1414 environments.
1416 21. It appears that in at least one Mac OS environment, the isxdigit() function
1417 is implemented as a macro that evaluates to its argument more than once,
1418 contravening the C 90 Standard (I haven't checked a later standard). There
1419 was an instance in pcretest which caused it to go wrong when processing
1420 \x{...} escapes in subject strings. The has been rewritten to avoid using
1421 things like p++ in the argument of isxdigit().
1424 Version 8.13 16-Aug-2011
1425 ------------------------
1427 1. The Unicode data tables have been updated to Unicode 6.0.0.
1429 2. Two minor typos in pcre_internal.h have been fixed.
1431 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1432 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1433 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1435 4. There were a number of related bugs in the code for matching backrefences
1436 caselessly in UTF-8 mode when codes for the characters concerned were
1437 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1438 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1439 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1440 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1441 2-byte code at the end of the subject (it thought there wasn't enough data
1442 left).
1444 5. Comprehensive information about what went wrong is now returned by
1445 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1446 as the output vector has at least 2 elements. The offset of the start of
1447 the failing character and a reason code are placed in the vector.
1449 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1450 now returned is for the first byte of the failing character, instead of the
1451 last byte inspected. This is an incompatible change, but I hope it is small
1452 enough not to be a problem. It makes the returned offset consistent with
1453 pcre_exec() and pcre_dfa_exec().
1455 7. pcretest now gives a text phrase as well as the error number when
1456 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1457 failure, the offset and reason code are output.
1459 8. When \R was used with a maximizing quantifier it failed to skip backwards
1460 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1461 back over a single character (\n). This seems wrong (because it treated the
1462 two characters as a single entity when going forwards), conflicts with the
1463 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1464 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1465 has been changed.
1467 9. Some internal refactoring has changed the processing so that the handling
1468 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1469 time (the PCRE_DOTALL option was changed this way some time ago: version
1470 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1471 which was always a bit of a fudge. It also means that there is one less
1472 argument for the match() function, which reduces its stack requirements
1473 slightly. This change also fixes an incompatibility with Perl: the pattern
1474 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1476 10. More internal refactoring has drastically reduced the number of recursive
1477 calls to match() for possessively repeated groups such as (abc)++ when
1478 using pcre_exec().
1480 11. While implementing 10, a number of bugs in the handling of groups were
1481 discovered and fixed:
1483 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1484 (a|)*(?1) gave a compile-time internal error.
1485 ((a|)+)+ did not notice that the outer group could match an empty string.
1486 (^a|^)+ was not marked as anchored.
1487 (.*a|.*)+ was not marked as matching at start or after a newline.
1489 12. Yet more internal refactoring has removed another argument from the match()
1490 function. Special calls to this function are now indicated by setting a
1491 value in a variable in the "match data" data block.
1493 13. Be more explicit in pcre_study() instead of relying on "default" for
1494 opcodes that mean there is no starting character; this means that when new
1495 ones are added and accidentally left out of pcre_study(), testing should
1496 pick them up.
1498 14. The -s option of pcretest has been documented for ages as being an old
1499 synonym of -m (show memory usage). I have changed it to mean "force study
1500 for every regex", that is, assume /S for every regex. This is similar to -i
1501 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1502 using it. It makes it easier to run collections of tests with and without
1503 study enabled, and thereby test pcre_study() more easily. All the standard
1504 tests are now run with and without -s (but some patterns can be marked as
1505 "never study" - see 20 below).
1507 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1508 restoration of the capturing data to the outer values was not happening
1509 correctly.
1511 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1512 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1513 pattern had matched an empty string, and so incorrectly returned a no
1514 match.
1516 17. There was optimizing code for the last branch of non-capturing parentheses,
1517 and also for the obeyed branch of a conditional subexpression, which used
1518 tail recursion to cut down on stack usage. Unfortunately, now that there is
1519 the possibility of (*THEN) occurring in these branches, tail recursion is
1520 no longer possible because the return has to be checked for (*THEN). These
1521 two optimizations have therefore been removed. [But see 8.20/11 above.]
1523 18. If a pattern containing \R was studied, it was assumed that \R always
1524 matched two bytes, thus causing the minimum subject length to be
1525 incorrectly computed because \R can also match just one byte.
1527 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1528 was incorrectly computed.
1530 20. If /S is present twice on a test pattern in pcretest input, it now
1531 *disables* studying, thereby overriding the use of -s on the command line
1532 (see 14 above). This is necessary for one or two tests to keep the output
1533 identical in both cases.
1535 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1536 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1538 22. When an atomic group that contained a capturing parenthesis was
1539 successfully matched, but the branch in which it appeared failed, the
1540 capturing was not being forgotten if a higher numbered group was later
1541 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1542 group 1 to "a", when in fact it should be unset. This applied to multi-
1543 branched capturing and non-capturing groups, repeated or not, and also to
1544 positive assertions (capturing in negative assertions does not happen
1545 in PCRE) and also to nested atomic groups.
1547 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1548 subject after a captured substring, to make it easier to tell which of a
1549 number of identical substrings has been captured.
1551 24. The way atomic groups are processed by pcre_exec() has been changed so that
1552 if they are repeated, backtracking one repetition now resets captured
1553 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1554 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1555 "aaa". Previously, it would have been "a". As part of this code
1556 refactoring, the way recursive calls are handled has also been changed.
1558 25. If an assertion condition captured any substrings, they were not passed
1559 back unless some other capturing happened later. For example, if
1560 (?(?=(a))a) was matched against "a", no capturing was returned.
1562 26. When studying a pattern that contained subroutine calls or assertions,
1563 the code for finding the minimum length of a possible match was handling
1564 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1565 group 1 called group 2 while simultaneously a separate group 2 called group
1566 1). A stack overflow occurred in this case. I have fixed this by limiting
1567 the recursion depth to 10.
1569 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1570 Fortmann. This supports explicit test numbers on the command line, and has
1571 argument validation and error reporting.
1573 28. An instance of \X with an unlimited repeat could fail if at any point the
1574 first character it looked at was a mark character.
1576 29. Some minor code refactoring concerning Unicode properties and scripts
1577 should reduce the stack requirement of match() slightly.
1579 30. Added the '=' option to pcretest to check the setting of unused capturing
1580 slots at the end of the pattern, which are documented as being -1, but are
1581 not included in the return count.
1583 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1584 compiled something random. Now it gives a compile-time error (as does
1585 Perl).
1587 32. A *MARK encountered during the processing of a positive assertion is now
1588 recorded and passed back (compatible with Perl).
1590 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1591 had alternative anchored branches, the search for a second match in a line
1592 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1593 matched the line "0102" twice. The same bug affected patterns that started
1594 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1595 twice.
1597 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1598 does, and because of capturing effects, quantifying parenthesized
1599 assertions may at times be useful. Quantifiers are now allowed for
1600 parenthesized assertions.
1602 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1604 36. \g was being checked for fancy things in a character class, when it should
1605 just be a literal "g".
1607 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1608 appearance of a nested POSIX class supersedes an apparent external class.
1609 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1610 unescaped square brackets may also appear as part of class names. For
1611 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1612 more like Perl. (But see 8.20/1 above.)
1614 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1615 was because it thought it was \N{name}, which is not supported).
1617 39. Add minix to OS list not supporting the -S option in pcretest.
1619 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1620 cannot analyze patterns in sufficient detail to catch mutual recursions
1621 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1622 subgroup is called recursively as a subpattern for a second time at the
1623 same position in the subject string. In previous releases this might have
1624 been caught by the recursion limit, or it might have run out of stack.
1626 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1627 happen only once. PCRE was, however incorrectly giving a compile time error
1628 "recursive call could loop indefinitely" because it cannot analyze the
1629 pattern in sufficient detail. The compile time test no longer happens when
1630 PCRE is compiling a conditional subpattern, but actual runaway loops are
1631 now caught at runtime (see 40 above).
1633 42. It seems that Perl allows any characters other than a closing parenthesis
1634 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1635 has been changed to be the same.
1637 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1638 as not to get warnings when autogen.sh is called. Also changed
1639 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1641 44. To help people who use pcregrep to scan files containing exceedingly long
1642 lines, the following changes have been made:
1644 (a) The default value of the buffer size parameter has been increased from
1645 8K to 20K. (The actual buffer used is three times this size.)
1647 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1648 PCRE is built.
1650 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1651 to be set at run time.
1653 (d) Numerical values in pcregrep options can be followed by K or M, for
1654 example --buffer-size=50K.
1656 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1657 given and the return code is set to 2.
1659 45. Add a pointer to the latest mark to the callout data block.
1661 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1662 partial match of an empty string instead of no match. This was specific to
1663 the use of ".".
1665 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1666 complete match instead of a partial match. This bug was dependent on both
1667 the PCRE_UTF8 and PCRE_DOTALL options being set.
1669 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1670 starting byte set, because \b was not being ignored.
1673 Version 8.12 15-Jan-2011
1674 ------------------------
1676 1. Fixed some typos in the markup of the man pages, and wrote a script that
1677 checks for such things as part of the documentation building process.
1679 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1680 --match-limit and --recursion-limit options (added for 8.11). In
1681 particular, this made one of the standard tests fail. (The integer value
1682 went into the wrong half of a long int.)
1684 3. If the --colour option was given to pcregrep with -v (invert match), it
1685 did strange things, either producing crazy output, or crashing. It should,
1686 of course, ignore a request for colour when reporting lines that do not
1687 match.
1689 4. Another pcregrep bug caused similar problems if --colour was specified with
1690 -M (multiline) and the pattern match finished with a line ending.
1692 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1693 matched in multiline mode, the following line was shown as part of the
1694 match. This seems wrong, so I have changed it.
1696 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1697 the check for further matches in the same line (so they could be coloured)
1698 to overrun the end of the current line. If another match was found, it was
1699 incorrectly shown (and then shown again when found in the next line).
1701 7. If pcregrep was compiled under Windows, there was a reference to the
1702 function pcregrep_exit() before it was defined. I am assuming this was
1703 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1704 reported by a user. I've moved the definition above the reference.
1707 Version 8.11 10-Dec-2010
1708 ------------------------
1710 1. (*THEN) was not working properly if there were untried alternatives prior
1711 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1712 backtracked to try for "b" instead of moving to the next alternative branch
1713 at the same level (in this case, to look for "c"). The Perl documentation
1714 is clear that when (*THEN) is backtracked onto, it goes to the "next
1715 alternative in the innermost enclosing group".
1717 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1718 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1719 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1720 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1721 (*THEN).
1723 3. If \s appeared in a character class, it removed the VT character from
1724 the class, even if it had been included by some previous item, for example
1725 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1726 of \s, but is part of the POSIX "space" class.)
1728 4. A partial match never returns an empty string (because you can always
1729 match an empty string at the end of the subject); however the checking for
1730 an empty string was starting at the "start of match" point. This has been
1731 changed to the "earliest inspected character" point, because the returned
1732 data for a partial match starts at this character. This means that, for
1733 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1734 (previously it gave "no match").
1736 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1737 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1738 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1739 has an implication that the given string is incomplete (because a partial
1740 match is preferred over a full match). For this reason, these items now
1741 give a partial match in this situation. [Aside: previously, the one case
1742 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1743 match rather than a full match, which was wrong by the old rules, but is
1744 now correct.]
1746 6. There was a bug in the handling of #-introduced comments, recognized when
1747 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1748 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1749 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1750 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1751 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1752 places in pcre_compile().
1754 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1755 comments when looking ahead for named forward references to subpatterns,
1756 the only newline sequence it recognized was NL. It now handles newlines
1757 according to the set newline convention.
1759 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1760 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1761 cater for a lack of strerror(). These oversights have been fixed.
1763 9. Added --match-limit and --recursion-limit to pcregrep.
1765 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1767 11. When the -o option was used, pcregrep was setting a return code of 1, even
1768 when matches were found, and --line-buffered was not being honoured.
1770 12. Added an optional parentheses number to the -o and --only-matching options
1771 of pcregrep.
1773 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1774 can match an empty string. The code to do it in pcretest and pcredemo
1775 needed fixing:
1777 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1778 only one byte after an empty string match just before CRLF (this case
1779 just got forgotten; "any" and "anycrlf" were OK).
1781 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1782 mode when an empty string match preceded an ASCII character followed by
1783 a non-ASCII character. (The code for advancing by one character rather
1784 than one byte was nonsense.)
1786 (c) The pcredemo.c sample program did not have any code at all to handle
1787 the cases when CRLF is a valid newline sequence.
1789 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1790 as a starting offset was within the subject string. There is now a new
1791 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1792 negative or greater than the length of the string. In order to test this,
1793 pcretest is extended to allow the setting of negative starting offsets.
1795 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1796 starting offset points to the beginning of a UTF-8 character was
1797 unnecessarily clumsy. I tidied it up.
1799 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1800 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1802 17. Nobody had reported that the --include_dir option, which was added in
1803 release 7.7 should have been called --include-dir (hyphen, not underscore)
1804 for compatibility with GNU grep. I have changed it to --include-dir, but
1805 left --include_dir as an undocumented synonym, and the same for
1806 --exclude-dir, though that is not available in GNU grep, at least as of
1807 release 2.5.4.
1809 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1810 characters from a string of bytes) have been redefined so as not to use
1811 loops, in order to improve performance in some environments. At the same
1812 time, I abstracted some of the common code into auxiliary macros to save
1813 repetition (this should not affect the compiled code).
1815 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1816 compile-time error is now given if \c is not followed by an ASCII
1817 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1818 different, and any byte value is allowed.)
1820 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1821 START_OPTIMIZE option, which is now allowed at compile time - but just
1822 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1823 to pcregrep and other applications that have no direct access to PCRE
1824 options. The new /Y option in pcretest sets this option when calling
1825 pcre_compile().
1827 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1828 back references. Groups containing recursive back references were forced to
1829 be atomic by that change, but in the case of named groups, the amount of
1830 memory required was incorrectly computed, leading to "Failed: internal
1831 error: code overflow". This has been fixed.
1833 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1834 pcretest.c, to avoid build problems in some Borland environments.
1837 Version 8.10 25-Jun-2010
1838 ------------------------
1840 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1841 THEN.
1843 2. (*ACCEPT) was not working when inside an atomic group.
1845 3. Inside a character class, \B is treated as a literal by default, but
1846 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1847 causes the error). The code is unchanged, but I tidied the documentation.
1849 4. Inside a character class, PCRE always treated \R and \X as literals,
1850 whereas Perl faults them if its -w option is set. I have changed PCRE so
1851 that it faults them when PCRE_EXTRA is set.
1853 5. Added support for \N, which always matches any character other than
1854 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1856 6. When compiling pcregrep with newer versions of gcc which may have
1857 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1858 declared with attribute warn_unused_result" were given. Just casting the
1859 result to (void) does not stop the warnings; a more elaborate fudge is
1860 needed. I've used a macro to implement this.
1862 7. Minor change to pcretest.c to avoid a compiler warning.
1864 8. Added four artifical Unicode properties to help with an option to make
1865 \s etc use properties (see next item). The new properties are: Xan
1866 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1868 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1869 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1870 this option. Modified pcretest to add /W to test this facility. Added
1871 REG_UCP to make it available via the POSIX interface.
1873 10. Added --line-buffered to pcregrep.
1875 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1876 studied, and the match started with a letter with a code point greater than
1877 127 whose first byte was different to the first byte of the other case of
1878 the letter, the other case of this starting letter was not recognized
1879 (#976).
1881 12. If a pattern that was studied started with a repeated Unicode property
1882 test, for example, \p{Nd}+, there was the theoretical possibility of
1883 setting up an incorrect bitmap of starting bytes, but fortunately it could
1884 not have actually happened in practice until change 8 above was made (it
1885 added property types that matched character-matching opcodes).
1887 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1888 possible starting bytes for non-anchored patterns.
1890 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1891 \R, and also a number of cases that involve Unicode properties, both
1892 explicit and implicit when PCRE_UCP is set.
1894 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1895 input, it could crash or give wrong results if characters with values
1896 greater than 0xc0 were present in the subject string. (Detail: it assumed
1897 UTF-8 input when processing these items.)
1899 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1900 size_t is 64-bit (#991).
1902 17. Added a check for running out of memory when PCRE is compiled with
1903 --disable-stack-for-recursion (#990).
1905 18. If the last data line in a file for pcretest does not have a newline on
1906 the end, a newline was missing in the output.
1908 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1909 less than 128) in its various bitmaps. However, there is a facility for
1910 generating tables according to the current locale when PCRE is compiled. It
1911 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1912 characters, are recognized by isspace() and therefore were getting set in
1913 these tables, and indeed these tables seem to approximate to ISO 8859. This
1914 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1915 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1916 which of course cannot start UTF-8 characters. I have changed the code so
1917 that only real ASCII characters (less than 128) and the correct starting
1918 bytes for UTF-8 encodings are set for characters greater than 127 when in
1919 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1920 altogether.)
1922 20. Added the /T option to pcretest so as to be able to run tests with non-
1923 standard character tables, thus making it possible to include the tests
1924 used for 19 above in the standard set of tests.
1926 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1927 reference to a subpattern the other side of a comment that contains an
1928 opening parenthesis caused either an internal compiling error, or a
1929 reference to the wrong subpattern.
1932 Version 8.02 19-Mar-2010
1933 ------------------------
1935 1. The Unicode data tables have been updated to Unicode 5.2.0.
1937 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1938 configured.
1940 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1941 original author of that file, following a query about its status.
1943 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1944 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1946 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1947 quantifier applied to a forward-referencing subroutine call, could compile
1948 incorrect code or give the error "internal error: previously-checked
1949 referenced subpattern not found".
1951 6. Both MS Visual Studio and Symbian OS have problems with initializing
1952 variables to point to external functions. For these systems, therefore,
1953 pcre_malloc etc. are now initialized to local functions that call the
1954 relevant global functions.
1956 7. There were two entries missing in the vectors called coptable and poptable
1957 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1958 I've fixed the data, and added a kludgy way of testing at compile time that
1959 the lengths are correct (equal to the number of opcodes).
1961 8. Following on from 7, I added a similar kludge to check the length of the
1962 eint vector in pcreposix.c.
1964 9. Error texts for pcre_compile() are held as one long string to avoid too
1965 much relocation at load time. To find a text, the string is searched,
1966 counting zeros. There was no check for running off the end of the string,
1967 which could happen if a new error number was added without updating the
1968 string.
1970 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1972 11. \K was not working if it appeared in an atomic group or in a group that
1973 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1974 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1975 the assertion is positive, but not if it is negative.
1977 12. Change 11 fortuitously reduced the size of the stack frame used in the
1978 "match()" function of pcre_exec.c by one pointer. Forthcoming
1979 implementation of support for (*MARK) will need an extra pointer on the
1980 stack; I have reserved it now, so that the stack frame size does not
1981 decrease.
1983 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1984 item in branch that calls a recursion is a subroutine call - as in the
1985 second branch in the above example - was incorrectly given the compile-
1986 time error "recursive call could loop indefinitely" because pcre_compile()
1987 was not correctly checking the subroutine for matching a non-empty string.
1989 14. The checks for overrunning compiling workspace could trigger after an
1990 overrun had occurred. This is a "should never occur" error, but it can be
1991 triggered by pathological patterns such as hundreds of nested parentheses.
1992 The checks now trigger 100 bytes before the end of the workspace.
1994 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1997 Version 8.01 19-Jan-2010
1998 ------------------------
2000 1. If a pattern contained a conditional subpattern with only one branch (in
2001 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2002 computed the wrong minimum data length (which is of course zero for such
2003 subpatterns). This could cause incorrect "no match" results.
2005 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2006 the pattern is reset in the first branch, pcre_compile() failed with
2007 "internal error: code overflow at offset...". This happened only when
2008 the reset was to the original external option setting. (An optimization
2009 abstracts leading options settings into an external setting, which was the
2010 cause of this.)
2012 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2013 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2014 assertion pattern did not match (meaning that the assertion was true), it
2015 was incorrectly treated as false if the SKIP had been reached during the
2016 matching. This also applied to assertions used as conditions.
2018 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2019 assertion subpattern, including such a pattern used as a condition,
2020 unpredictable results occurred, instead of the error return
2023 5. The C++ GlobalReplace function was not working like Perl for the special
2024 situation when an empty string is matched. It now does the fancy magic
2025 stuff that is necessary.
2027 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2028 removed. (These were left over from very, very early versions of PCRE.)
2030 7. Some cosmetic changes to the code to make life easier when compiling it
2031 as part of something else:
2033 (a) Change DEBUG to PCRE_DEBUG.
2035 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2036 called "current" as "current_branch", to prevent a collision with the
2037 Linux macro when compiled as a kernel module.
2039 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2040 prevent a collision with the Linux macro when compiled as a kernel
2041 module.
2043 8. In pcre_compile() there are some checks for integer overflows that used to
2044 cast potentially large values to (double). This has been changed to that
2045 when building, a check for int64_t is made, and if it is found, it is used
2046 instead, thus avoiding the use of floating point arithmetic. (There is no
2047 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2048 double.
2050 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2051 2005 (difference between two addresses compared to an unsigned value).
2053 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2054 custom one, because of the following reported problem in Windows:
2056 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2057 under Win32.
2058 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2059 therefore missing the function definition.
2060 - The compiler thus generates a "C" signature for the test function.
2061 - The linker fails to find the "C" function.
2062 - PCRE fails to configure if asked to do so against libbz2.
2064 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2065 messages were output:
2067 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2068 rerunning libtoolize, to keep the correct libtool macros in-tree.
2069 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2071 I have done both of these things.
2073 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2074 most of the time, it *can* run out if it is given a pattern that contains a
2075 runaway infinite recursion. I updated the discussion in the pcrestack man
2076 page.
2078 13. Now that we have gone to the x.xx style of version numbers, the minor
2079 version may start with zero. Using 08 or 09 is a bad idea because users
2080 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2081 interpreted as invalid octal numbers. I've updated the previous comment in
2082 configure.ac, and also added a check that gives an error if 08 or 09 are
2083 used.
2085 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2086 causing partial matching to fail when the end of the subject matched \W
2087 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2089 15. There were some discrepancies between the declarations in pcre_internal.h
2090 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2091 their definitions. The declarations used "const uschar *" and the
2092 definitions used USPTR. Even though USPTR is normally defined as "const
2093 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2094 reported that: "This difference in casting confuses some C++ compilers, for
2095 example, SunCC recognizes above declarations as different functions and
2096 generates broken code for hbpcre." I have changed the declarations to use
2097 USPTR.
2099 16. GNU libtool is named differently on some systems. The autogen.sh script now
2100 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2101 (FreeBSD).
2103 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2104 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2105 comment: "Figure out how to create a longlong from a string: strtoll and
2106 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2107 instance, but it only takes 2 args instead of 3!"
2109 18. A subtle bug concerned with back references has been fixed by a change of
2110 specification, with a corresponding code fix. A pattern such as
2111 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2112 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2113 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2114 same bug. Such groups have to be quantified to be useful, or contained
2115 inside another quantified group. (If there's no repetition, the reference
2116 can never match.) The problem arises because, having left the group and
2117 moved on to the rest of the pattern, a later failure that backtracks into
2118 the group uses the captured value from the final iteration of the group
2119 rather than the correct earlier one. I have fixed this in PCRE by forcing
2120 any group that contains a reference to itself to be an atomic group; that
2121 is, there cannot be any backtracking into it once it has completed. This is
2122 similar to recursive and subroutine calls.
2125 Version 8.00 19-Oct-09
2126 ----------------------
2128 1. The table for translating pcre_compile() error codes into POSIX error codes
2129 was out-of-date, and there was no check on the pcre_compile() error code
2130 being within the table. This could lead to an OK return being given in
2131 error.
2133 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2134 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2135 in a Windows environment.
2137 3. The pcregrep --count option prints the count for each file even when it is
2138 zero, as does GNU grep. However, pcregrep was also printing all files when
2139 --files-with-matches was added. Now, when both options are given, it prints
2140 counts only for those files that have at least one match. (GNU grep just
2141 prints the file name in this circumstance, but including the count seems
2142 more useful - otherwise, why use --count?) Also ensured that the
2143 combination -clh just lists non-zero counts, with no names.
2145 4. The long form of the pcregrep -F option was incorrectly implemented as
2146 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2147 but it seems right to fix it, and I didn't think it was worth preserving
2148 the old behaviour.
2150 5. The command line items --regex=pattern and --regexp=pattern were not
2151 recognized by pcregrep, which required --regex pattern or --regexp pattern
2152 (with a space rather than an '='). The man page documented the '=' forms,
2153 which are compatible with GNU grep; these now work.
2155 6. No libpcreposix.pc file was created for pkg-config; there was just
2156 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2158 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2159 when UCP support is not needed, by modifying the Python script that
2160 generates it from Unicode data files. This should not matter if the module
2161 is correctly used as a library, but I received one complaint about 50K of
2162 unwanted data. My guess is that the person linked everything into his
2163 program rather than using a library. Anyway, it does no harm.
2165 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2166 was a minimum greater than 1 for a wide character in a possessive
2167 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2168 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2169 character. Chaos in the form of incorrect output or a compiling loop could
2170 result.
2172 9. The restrictions on what a pattern can contain when partial matching is
2173 requested for pcre_exec() have been removed. All patterns can now be
2174 partially matched by this function. In addition, if there are at least two
2175 slots in the offset vector, the offset of the earliest inspected character
2176 for the match and the offset of the end of the subject are set in them when
2177 PCRE_ERROR_PARTIAL is returned.
2179 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2180 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2181 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2182 and may be more useful for multi-segment matching.
2184 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2185 used to be given if ever the end of the subject was reached; now it is
2186 given only if matching could not proceed because another character was
2187 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2188 string "Z", which now yields "no match" instead of "partial match". In the
2189 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2190 final character ended with (*FAIL).
2192 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2193 if the pattern had a "must contain" character that was already found in the
2194 earlier partial match, unless partial matching was again requested. For
2195 example, with the pattern /dog.(body)?/, the "must contain" character is
2196 "g". If the first part-match was for the string "dog", restarting with
2197 "sbody" failed. This bug has been fixed.
2199 13. The string returned by pcre_dfa_exec() after a partial match has been
2200 changed so that it starts at the first inspected character rather than the
2201 first character of the match. This makes a difference only if the pattern
2202 starts with a lookbehind assertion or \b or \B (\K is not supported by
2203 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2204 matching functions compatible, and I think it's the right thing to do.
2206 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2207 so that the demonstration program is easily available in environments where
2208 PCRE has not been installed from source.
2210 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2211 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2212 library.
2214 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2215 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2216 is not the first non-POSIX option to be added. Clearly some people find
2217 these options useful.
2219 17. If a caller to the POSIX matching function regexec() passes a non-zero
2220 value for nmatch with a NULL value for pmatch, the value of
2221 nmatch is forced to zero.
2223 18. RunGrepTest did not have a test for the availability of the -u option of
2224 the diff command, as RunTest does. It now checks in the same way as
2225 RunTest, and also checks for the -b option.
2227 19. If an odd number of negated classes containing just a single character
2228 interposed, within parentheses, between a forward reference to a named
2229 subpattern and the definition of the subpattern, compilation crashed with
2230 an internal error, complaining that it could not find the referenced
2231 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2232 [The bug was that it was starting one character too far in when skipping
2233 over the character class, thus treating the ] as data rather than
2234 terminating the class. This meant it could skip too much.]
2236 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2237 /g option in pcretest when the pattern contains \K, which makes it possible
2238 to have an empty string match not at the start, even when the pattern is
2239 anchored. Updated pcretest and pcredemo to use this option.
2241 21. If the maximum number of capturing subpatterns in a recursion was greater
2242 than the maximum at the outer level, the higher number was returned, but
2243 with unset values at the outer level. The correct (outer level) value is
2244 now given.
2246 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2247 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2248 make it do so. The string so far is captured, making this feature
2249 compatible with Perl.
2251 23. The tests have been re-organized, adding tests 11 and 12, to make it
2252 possible to check the Perl 5.10 features against Perl 5.10.
2254 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2255 pattern matches a fixed length string. PCRE did not allow this; now it
2256 does. Neither allows recursion.
2258 25. I finally figured out how to implement a request to provide the minimum
2259 length of subject string that was needed in order to match a given pattern.
2260 (It was back references and recursion that I had previously got hung up
2261 on.) This code has now been added to pcre_study(); it finds a lower bound
2262 to the length of subject needed. It is not necessarily the greatest lower
2263 bound, but using it to avoid searching strings that are too short does give
2264 some useful speed-ups. The value is available to calling programs via
2265 pcre_fullinfo().
2267 26. While implementing 25, I discovered to my embarrassment that pcretest had
2268 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2269 study optimizations had never been tested with that matching function.
2270 Oops. What is worse, even when it was passed study data, there was a bug in
2271 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2272 were also very few tests of studied patterns with pcre_dfa_exec().
2274 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2275 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2276 on the other side of the coin, they are no longer allowed to have different
2277 names, because these cannot be distinguished in PCRE, and this has caused
2278 confusion. (This is a difference from Perl.)
2280 28. When duplicate subpattern names are present (necessarily with different
2281 numbers, as required by 27 above), and a test is made by name in a
2282 conditional pattern, either for a subpattern having been matched, or for
2283 recursion in such a pattern, all the associated numbered subpatterns are
2284 tested, and the overall condition is true if the condition is true for any
2285 one of them. This is the way Perl works, and is also more like the way
2286 testing by number works.
2289 Version 7.9 11-Apr-09
2290 ---------------------
2292 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2293 (pcretest), all targets were linked against these libraries. This included
2294 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2295 libraries. This caused unwanted dependencies to be created. This problem
2296 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2297 pcretest is linked with readline.
2299 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2300 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2301 moved outside it again, because FALSE and TRUE are already defined in AIX,
2302 but BOOL is not.
2304 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2305 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2307 4. The pcregrep documentation said spaces were inserted as well as colons (or
2308 hyphens) following file names and line numbers when outputting matching
2309 lines. This is not true; no spaces are inserted. I have also clarified the
2310 wording for the --colour (or --color) option.
2312 5. In pcregrep, when --colour was used with -o, the list of matching strings
2313 was not coloured; this is different to GNU grep, so I have changed it to be
2314 the same.
2316 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2317 each matching line was coloured. Now it goes on to look for further matches
2318 of any of the test patterns, which is the same behaviour as GNU grep.
2320 7. A pattern that could match an empty string could cause pcregrep to loop; it
2321 doesn't make sense to accept an empty string match in pcregrep, so I have
2322 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2323 seems to be how GNU grep behaves. [But see later change 40 for release
2324 8.33.]
2326 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2327 start or after a newline", because the conditional assertion was not being
2328 correctly handled. The rule now is that both the assertion and what follows
2329 in the first alternative must satisfy the test.
2331 9. If auto-callout was enabled in a pattern with a conditional group whose
2332 condition was an assertion, PCRE could crash during matching, both with
2333 pcre_exec() and pcre_dfa_exec().
2335 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2336 used for matching.
2338 11. Unicode property support in character classes was not working for
2339 characters (bytes) greater than 127 when not in UTF-8 mode.
2341 12. Added the -M command line option to pcretest.
2343 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2345 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2347 16. Added comments and documentation about mis-use of no_arg in the C++
2348 wrapper.
2350 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2351 from Martin Jerabek that uses macro names for all relevant character and
2352 string constants.
2354 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2355 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2356 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2357 these, but not everybody uses configure.
2359 19. A conditional group that had only one branch was not being correctly
2360 recognized as an item that could match an empty string. This meant that an
2361 enclosing group might also not be so recognized, causing infinite looping
2362 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2363 with the subject "ab", where knowledge that the repeated group can match
2364 nothing is needed in order to break the loop.
2366 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2367 exec(), but without supplying a callout function, matching went wrong.
2369 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2370 leak if the size of the offset vector was greater than 30. When the vector
2371 is smaller, the saved offsets during recursion go onto a local stack
2372 vector, but for larger vectors malloc() is used. It was failing to free
2373 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2374 error, in fact).
2376 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2377 heapframe that is used only when UTF-8 support is enabled. This caused no
2378 problem, but was untidy.
2380 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2381 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2382 included within another project.
2384 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2385 slightly modified by me:
2387 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2388 not building pcregrep.
2390 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2391 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2393 25. Forward references, both numeric and by name, in patterns that made use of
2394 duplicate group numbers, could behave incorrectly or give incorrect errors,
2395 because when scanning forward to find the reference group, PCRE was not
2396 taking into account the duplicate group numbers. A pattern such as
2397 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2399 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2400 the feature of a custom pointer more persuasive (as requested by a user).
2402 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2403 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2404 pre-defined.
2406 28. Added support for (*UTF8) at the start of a pattern.
2408 29. Arrange for flags added by the "release type" setting in CMake to be shown
2409 in the configuration summary.
2412 Version 7.8 05-Sep-08
2413 ---------------------
2415 1. Replaced UCP searching code with optimized version as implemented for Ad
2416 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2417 stage table and inline lookup instead of a function, giving speed ups of 2
2418 to 5 times on some simple patterns that I tested. Permission was given to
2419 distribute the MultiStage2.py script that generates the tables (it's not in
2420 the tarball, but is in the Subversion repository).
2422 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2423 scripts.
2425 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2426 a group with a zero qualifier. The result of the study could be incorrect,
2427 or the function might crash, depending on the pattern.
2429 4. Caseless matching was not working for non-ASCII characters in back
2430 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2431 It now works when Unicode Property Support is available.
2433 5. In pcretest, an escape such as \x{de} in the data was always generating
2434 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2435 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2436 truncation.
2438 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2440 7. Added two (int) casts to pcregrep when printing the difference of two
2441 pointers, in case they are 64-bit values.
2443 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2444 test 2 if it fails.
2446 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2447 and a #define of that name to empty if it is not externally set. This is to
2448 allow users of MSVC to set it if necessary.
2450 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2451 the convenience functions in the pcre_get.c source file.
2453 11. An option change at the start of a pattern that had top-level alternatives
2454 could cause overwriting and/or a crash. This command provoked a crash in
2455 some environments:
2457 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2459 This potential security problem was recorded as CVE-2008-2371.
2461 12. For a pattern where the match had to start at the beginning or immediately
2462 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2463 pcre_dfa_exec() could read past the end of the passed subject if there was
2464 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2465 pcretest so that it places the subject at the end of its malloc-ed buffer.
2467 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2468 exec() might read past the end of the data buffer in UTF-8 mode.
2470 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2471 the data contained the byte 0x85 as part of a UTF-8 character within its
2472 first line. This applied both to normal and DFA matching.
2474 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2475 /^[^d]*?$/8 failed to match "abc".
2477 16. Added a missing copyright notice to pcrecpp_internal.h.
2479 17. Make it more clear in the documentation that values returned from
2480 pcre_exec() in ovector are byte offsets, not character counts.
2482 18. Tidied a few places to stop certain compilers from issuing warnings.
2484 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2485 supplied by Stefan Weber. I made a further small update for 7.8 because
2486 there is a change of source arrangements: the pcre_searchfuncs.c module is
2487 replaced by pcre_ucd.c.
2490 Version 7.7 07-May-08
2491 ---------------------
2493 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2494 a string to a long long, pretend we don't even have a long long." This is
2495 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2497 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2498 pre-7.6 versions, which defined a global no_arg variable instead of putting
2499 it in the RE class. (See also #8 below.)
2501 3. Remove a line of dead code, identified by coverity and reported by Nuno
2502 Lopes.
2504 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2506 (1) The include/exclude patterns were being applied to the whole pathnames
2507 of files, instead of just to the final components.
2509 (2) If there was more than one level of directory, the subdirectories were
2510 skipped unless they satisfied the include/exclude conditions. This is
2511 inconsistent with GNU grep (and could even be seen as contrary to the
2512 pcregrep specification - which I improved to make it absolutely clear).
2513 The action now is always to scan all levels of directory, and just
2514 apply the include/exclude patterns to regular files.
2516 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2517 --exclude_dir in the tests to avoid scanning .svn directories.
2519 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2520 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2521 doesn't support NULs in patterns.
2523 7. Added some missing "const"s to declarations of static tables in
2524 pcre_compile.c and pcre_dfa_exec.c.
2526 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2527 caused by fix #2 above. (Subsequently also a second patch to fix the
2528 first patch. And a third patch - this was a messy problem.)
2530 9. Applied Craig's patch to remove the use of push_back().
2532 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2533 matching function regexec().
2535 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2536 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2537 references. PCRE supports relative numbers with this syntax (I don't think
2538 Oniguruma does).
2540 12. Previously, a group with a zero repeat such as (...){0} was completely
2541 omitted from the compiled regex. However, this means that if the group
2542 was called as a subroutine from elsewhere in the pattern, things went wrong
2543 (an internal error was given). Such groups are now left in the compiled
2544 pattern, with a new opcode that causes them to be skipped at execution
2545 time.
2547 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2548 to the way PCRE behaves:
2550 (a) A lone ] character is dis-allowed (Perl treats it as data).
2552 (b) A back reference to an unmatched subpattern matches an empty string
2553 (Perl fails the current match path).
2555 (c) A data ] in a character class must be notated as \] because if the
2556 first data character in a class is ], it defines an empty class. (In
2557 Perl it is not possible to have an empty class.) The empty class []
2558 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2559 The negative empty class [^] matches any one character, independently
2560 of the DOTALL setting.
2562 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2563 non-existent subpattern following a character class starting with ']' and
2564 containing () gave an internal compiling error instead of "reference to
2565 non-existent subpattern". Fortunately, when the pattern did exist, the
2566 compiled code was correct. (When scanning forwards to check for the
2567 existence of the subpattern, it was treating the data ']' as terminating
2568 the class, so got the count wrong. When actually compiling, the reference
2569 was subsequently set up correctly.)
2571 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2572 it was being rejected as not supported by pcre_dfa_exec(), even though
2573 other assertions are supported. I have made pcre_dfa_exec() support
2574 (*FAIL).
2576 16. The implementation of 13c above involved the invention of a new opcode,
2577 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2578 cannot be changed at match time, I realized I could make a small
2579 improvement to matching performance by compiling OP_ALLANY instead of
2580 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2581 on the OP_ANY path.
2583 17. Compiling pcretest on Windows with readline support failed without the
2584 following two fixes: (1) Make the unistd.h include conditional on
2585 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2587 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2588 ncurses library to be included for pcretest when ReadLine support is
2589 requested, but also to allow for it to be overridden. This patch came from
2590 Daniel Bergström.
2592 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2593 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2594 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2595 spotting this.
2598 Version 7.6 28-Jan-08
2599 ---------------------
2601 1. A character class containing a very large number of characters with
2602 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2603 overflow.
2605 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2606 HAVE_LONG_LONG is not defined.
2608 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2609 bring them up to date and include new features. This patch includes:
2611 - Fixed PH's badly added libz and libbz2 support.
2612 - Fixed a problem with static linking.
2613 - Added pcredemo. [But later removed - see 7 below.]
2614 - Fixed dftables problem and added an option.
2615 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2617 - Added readline support for pcretest.
2618 - Added an listing of the option settings after cmake has run.
2620 4. A user submitted a patch to Makefile that makes it easy to create
2621 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2622 Makefile.am that cause it to include this special target, without
2623 affecting anything else. Note that the same mingw target plus all
2624 the other distribution libraries and programs are now supported
2625 when configuring with CMake (see 6 below) instead of with
2626 Configure/Make.
2628 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2629 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2630 exported in the Windows port". It has not yet been confirmed that the patch
2631 solves the problem, but it does no harm.
2633 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2634 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2635 with CMake, and also correct the comment about stack recursion.
2637 7. Remove the automatic building of pcredemo from the ./configure system and
2638 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2639 of a program that users should build themselves after PCRE is installed, so
2640 building it automatically is not really right. What is more, it gave
2641 trouble in some build environments.
2643 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2646 Version 7.5 10-Jan-08
2647 ---------------------
2649 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2650 values in parens when parsing an RE using the C++ wrapper."
2652 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2653 Characters greater than 255 were excluded from the class instead of being
2654 included.
2656 3. The same bug as (2) above applied to negated POSIX classes such as
2657 [:^space:].
2659 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2660 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2661 I have changed it.
2663 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2664 first named subpattern) and a construct such as (?&a) would reference the
2665 first named subpattern whose name started with "a" (in other words, the
2666 length check was missing). Both these problems are fixed. "Subpattern name
2667 expected" is now given for (?&) (a zero-length name), and this patch also
2668 makes it give the same error for \k'' (previously it complained that that
2669 was a reference to a non-existent subpattern).
2671 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2672 this is right because (?- can be followed by option settings as well as by
2673 digits. I have, however, made the messages clearer.
2675 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2676 than the number used in the conditional) now cause a compile-time error.
2677 This is actually not compatible with Perl, which accepts such patterns, but
2678 treats the conditional as always being FALSE (as PCRE used to), but it
2679 seems to me that giving a diagnostic is better.
2681 8. Change "alphameric" to the more common word "alphanumeric" in comments
2682 and messages.
2684 9. Fix two occurrences of "backslash" in comments that should have been
2685 "backspace".
2687 10. Remove two redundant lines of code that can never be obeyed (their function
2688 was moved elsewhere).
2690 11. The program that makes PCRE's Unicode character property table had a bug
2691 which caused it to generate incorrect table entries for sequences of
2692 characters that have the same character type, but are in different scripts.
2693 It amalgamated them into a single range, with the script of the first of
2694 them. In other words, some characters were in the wrong script. There were
2695 thirteen such cases, affecting characters in the following ranges:
2697 U+002b0 - U+002c1
2698 U+0060c - U+0060d
2699 U+0061e - U+00612
2700 U+0064b - U+0065e
2701 U+0074d - U+0076d
2702 U+01800 - U+01805
2703 U+01d00 - U+01d77
2704 U+01d9b - U+01dbf
2705 U+0200b - U+0200f
2706 U+030fc - U+030fe
2707 U+03260 - U+0327f
2708 U+0fb46 - U+0fbb1
2709 U+10450 - U+1049d
2711 12. The -o option (show only the matching part of a line) for pcregrep was not
2712 compatible with GNU grep in that, if there was more than one match in a
2713 line, it showed only the first of them. It now behaves in the same way as
2714 GNU grep.
2716 13. If the -o and -v options were combined for pcregrep, it printed a blank
2717 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2718 does the same. The return code can be used to tell if there were any
2719 non-matching lines.
2721 14. Added --file-offsets and --line-offsets to pcregrep.
2723 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2724 infinitely looping recursion. The bug was that positive lookaheads were not
2725 being skipped when checking for a possible empty match (negative lookaheads
2726 and both kinds of lookbehind were skipped).
2728 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2729 inclusion of <windows.h> to before rather than after the definition of
2730 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2732 17. Specifying a possessive quantifier with a specific limit for a Unicode
2733 character property caused pcre_compile() to compile bad code, which led at
2734 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2735 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2736 caused the error; without that there was no problem.
2738 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2740 19. Added --enable-pcretest-libreadline.
2742 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2743 RE::GlobalReplace(). As a result, the number of replacements returned was
2744 double what it should be. I removed one of the increments, but Craig sent a
2745 later patch that removed the other one (the right fix) and added unit tests
2746 that check the return values (which was not done before).
2748 21. Several CMake things:
2750 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2751 the names libpcre and libpcreposix, not just pcre and pcreposix.
2753 (2) The above change means that pcretest and pcregrep are now correctly
2754 linked with the newly-built libraries, not previously installed ones.
2758 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2759 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2760 UTF-8 newline character). The key issue is that the pattern starts .*;
2761 this means that the match must be either at the beginning, or after a
2762 newline. The bug was in the code for advancing after a failed match and
2763 checking that the new position followed a newline. It was not taking
2764 account of UTF-8 characters correctly.
2766 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2767 character classes. PCRE was not treating the sequence [:...:] as a
2768 character class unless the ... were all letters. Perl, however, seems to
2769 allow any characters between [: and :], though of course it rejects as
2770 unknown any "names" that contain non-letters, because all the known class
2771 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2772 for example, whereas PCRE did not - it did not recognize a POSIX character
2773 class. This seemed a bit dangerous, so the code has been changed to be
2774 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2775 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2776 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2777 Perl does, and where it didn't before.
2779 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2780 Windows environments %n is disabled by default.
2783 Version 7.4 21-Sep-07
2784 ---------------------
2786 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2787 means that a class such as [\s] counted as "explicit reference to CR or
2788 LF". That isn't really right - the whole point of the change was to try to
2789 help when there was an actual mention of one of the two characters. So now
2790 the change happens only if \r or \n (or a literal CR or LF) character is
2791 encountered.
2793 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2794 of both had grown to the point where there were only 3 bits left.
2795 Fortunately, there was spare space in the data structure, and so I have
2796 moved the internal flags into a new 16-bit field to free up more option
2797 bits.
2799 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2800 but did not set the internal JCHANGED flag - either of these is enough to
2801 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2802 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2803 start sets both bits.
2805 4. Added options (at build time, compile time, exec time) to change \R from
2806 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2808 5. doc/pcresyntax.html was missing from the distribution.
2810 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2811 compatibility, even though it is no longer used.
2813 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2814 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2815 windows.h file is present (where different names are used). [This was
2816 reversed later after testing - see 16 below.]
2818 8. Changed all #include <config.h> to #include "config.h". There were also
2819 some further <pcre.h> cases that I changed to "pcre.h".
2821 9. When pcregrep was used with the --colour option, it missed the line ending
2822 sequence off the lines that it output.
2824 10. It was pointed out to me that arrays of string pointers cause lots of
2825 relocations when a shared library is dynamically loaded. A technique of
2826 using a single long string with a table of offsets can drastically reduce
2827 these. I have refactored PCRE in four places to do this. The result is
2828 dramatic:
2830 Originally: 290
2831 After changing UCP table: 187
2832 After changing error message table: 43
2833 After changing table of "verbs" 36
2834 After changing table of Posix names 22
2836 Thanks to the folks working on Gregex for glib for this insight.
2838 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2839 unicode-properties was also set.
2841 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2843 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2844 checked only for CRLF.
2846 14. Added casts to pcretest.c to avoid compiler warnings.
2848 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2850 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2851 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2852 entirely. This removes changes made in 7 above.
2854 17. The CMake files have been updated, and there is now more information about
2855 building with CMake in the NON-UNIX-USE document.
2858 Version 7.3 28-Aug-07
2859 ---------------------
2861 1. In the rejigging of the build system that eventually resulted in 7.1, the
2862 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2863 brackets there is not right, since it causes compilers to look for an
2864 installed pcre.h, not the version that is in the source that is being
2865 compiled (which of course may be different). I have changed it back to:
2867 #include "pcre.h"
2869 I have a vague recollection that the change was concerned with compiling in
2870 different directories, but in the new build system, that is taken care of
2871 by the VPATH setting the Makefile.
2873 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2874 when the subject happened to end in the byte 0x85 (e.g. if the last
2875 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2876 characters but of course it shouldn't be taken as a newline when it is part
2877 of another character. The bug was that, for an unlimited repeat of . in
2878 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2879 characters when looking for a newline.
2881 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2883 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2884 in debug output.
2886 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2887 long printing in the pcrecpp unittest when running under MinGW.
2889 6. ESC_K was left out of the EBCDIC table.
2891 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2892 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2893 limit also applies to "virtual nesting" when a pattern is recursive, and in
2894 this case 1000 isn't so big. I have been able to remove this limit at the
2895 expense of backing off one optimization in certain circumstances. Normally,
2896 when pcre_exec() would call its internal match() function recursively and
2897 immediately return the result unconditionally, it uses a "tail recursion"
2898 feature to save stack. However, when a subpattern that can match an empty
2899 string has an unlimited repetition quantifier, it no longer makes this
2900 optimization. That gives it a stack frame in which to save the data for
2901 checking that an empty string has been matched. Previously this was taken
2902 from the 1000-entry workspace that had been reserved. So now there is no
2903 explicit limit, but more stack is used.
2905 8. Applied Daniel's patches to solve problems with the import/export magic
2906 syntax that is required for Windows, and which was going wrong for the
2907 pcreposix and pcrecpp parts of the library. These were overlooked when this
2908 problem was solved for the main library.
2910 9. There were some crude static tests to avoid integer overflow when computing
2911 the size of patterns that contain repeated groups with explicit upper
2912 limits. As the maximum quantifier is 65535, the maximum group length was
2913 set at 30,000 so that the product of these two numbers did not overflow a
2914 32-bit integer. However, it turns out that people want to use groups that
2915 are longer than 30,000 bytes (though not repeat them that many times).
2916 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2917 made it possible to implement the integer overflow checks in a much more
2918 dynamic way, which I have now done. The artificial limitation on group
2919 length has been removed - we now have only the limit on the total length of
2920 the compiled pattern, which depends on the LINK_SIZE setting.
2922 10. Fixed a bug in the documentation for get/copy named substring when
2923 duplicate names are permitted. If none of the named substrings are set, the
2924 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2925 empty string.
2927 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2928 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2929 because the ] is interpreted as the first data character and the
2930 terminating ] is not found. PCRE has been made compatible with Perl in this
2931 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2932 cause memory overwriting.
2934 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2935 string has been matched (to stop an infinite loop). It was not recognizing
2936 a conditional subpattern that could match an empty string if that
2937 subpattern was within another subpattern. For example, it looped when
2938 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2939 condition was not nested. This bug has been fixed.
2941 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2942 past the start of the subject in the presence of bytes with the top bit
2943 set, for example "\x8aBCD".
2945 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2946 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2948 14. Optimized (?!) to (*FAIL).
2950 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2951 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2952 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2953 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2954 does: it's just the validity check that is more restrictive.
2956 16. Inserted checks for integer overflows during escape sequence (backslash)
2957 processing, and also fixed erroneous offset values for syntax errors during
2958 backslash processing.
2960 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2961 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2963 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2964 caused an overrun.
2966 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2967 something other than just ASCII characters) inside a group that had an
2968 unlimited repeat caused a loop at compile time (while checking to see
2969 whether the group could match an empty string).
2971 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2972 [\P{Any}] did so. (Error in the code for printing property names.)
2974 21. An orphan \E inside a character class could cause a crash.
2976 22. A repeated capturing bracket such as (A)? could cause a wild memory
2977 reference during compilation.
2979 23. There are several functions in pcre_compile() that scan along a compiled
2980 expression for various reasons (e.g. to see if it's fixed length for look
2981 behind). There were bugs in these functions when a repeated \p or \P was
2982 present in the pattern. These operators have additional parameters compared
2983 with \d, etc, and these were not being taken into account when moving along
2984 the compiled data. Specifically:
2986 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2987 length.
2989 (b) An item such as \pL+ within a repeated group could cause crashes or
2990 loops.
2992 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2993 "reference to non-existent subpattern" error.
2995 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2997 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2998 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3000 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3002 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3003 character were causing crashes (broken optimization).
3005 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3006 \p or \P) caused a compile-time loop.
3008 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3009 break. For example, the unstudied pattern [\r\n]A does not match the string
3010 "\r\nA" because change 7.0/46 below moves the current point on by two
3011 characters after failing to match at the start. However, the pattern \nA
3012 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3013 the same is true. There doesn't seem any very clean way out of this, but
3014 what I have chosen to do makes the common cases work: PCRE now takes note
3015 of whether there can be an explicit match for \r or \n anywhere in the
3016 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3017 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3018 pattern has explicit CR or LF references.
3020 29. Added (*CR) etc for changing newline setting at start of pattern.
3023 Version 7.2 19-Jun-07
3024 ---------------------
3026 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3027 which is apparently normally available under Windows.
3029 2. Re-jig the pcregrep tests with different newline settings in an attempt
3030 to make them independent of the local environment's newline setting.
3032 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3034 4. Some of the "internals" tests were previously cut out when the link size
3035 was not 2, because the output contained actual offsets. The recent new
3036 "Z" feature of pcretest means that these can be cut out, making the tests
3037 usable with all link sizes.
3039 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3040 stack recursion. This gives a massive performance boost under BSD, but just
3041 a small improvement under Linux. However, it saves one field in the frame
3042 in all cases.
3044 6. Added more features from the forthcoming Perl 5.10:
3046 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3047 recursion call. It refers to the nth most recently opened parentheses.
3049 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3050 to be opened parentheses.
3052 (c) Conditions that refer to capturing parentheses can be specified
3053 relatively, for example, (?(-2)... or (?(+3)...
3055 (d) \K resets the start of the current match so that everything before
3056 is not part of it.
3058 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3060 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3061 reference syntax.
3063 (g) (?| introduces a group in which the numbering of parentheses in each
3064 alternative starts with the same number.
3066 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3068 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3071 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3072 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3073 for detecting groups that can match an empty string.
3075 9. A pattern with a very large number of alternatives (more than several
3076 hundred) was running out of internal workspace during the pre-compile
3077 phase, where pcre_compile() figures out how much memory will be needed. A
3078 bit of new cunning has reduced the workspace needed for groups with
3079 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3080 workspace instead of running out of the 4096 that are available.
3082 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3084 11. Applied patch from Google to remove an optimization that didn't quite work.
3085 The report of the bug said:
3087 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3088 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3089 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3091 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3092 it matched the wrong number of bytes.
3095 Version 7.1 24-Apr-07
3096 ---------------------
3098 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3099 that is more "standard", making use of automake and other Autotools. There
3100 is some re-arrangement of the files and adjustment of comments consequent
3101 on this.
3103 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3104 for recursive directory scanning broke on some systems because the files
3105 are not scanned in any specific order and on different systems the order
3106 was different. A call to "sort" has been inserted into RunGrepTest for the
3107 approprate test as a short-term fix. In the longer term there may be an
3108 alternative.
3110 3. I had an email from Eric Raymond about problems translating some of PCRE's
3111 man pages to HTML (despite the fact that I distribute HTML pages, some
3112 people do their own conversions for various reasons). The problems
3113 concerned the use of low-level troff macros .br and .in. I have therefore
3114 removed all such uses from the man pages (some were redundant, some could
3115 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3116 HTML has been updated to handle .nf/.fi and to complain if it encounters
3117 .br or .in.
3119 4. Updated comments in configure.ac that get placed in config.h.in and also
3120 arranged for config.h to be included in the distribution, with the name
3121 config.h.generic, for the benefit of those who have to compile without
3122 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3124 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3125 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3126 makevp.bat for the current PCRE, using the additional files
3127 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3129 6. A Windows user reported a minor discrepancy with test 2, which turned out
3130 to be caused by a trailing space on an input line that had got lost in his
3131 copy. The trailing space was an accident, so I've just removed it.
3133 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3134 that is needed.
3136 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3137 as "const" (a) because they are and (b) because it helps the PHP
3138 maintainers who have recently made a script to detect big data structures
3139 in the php code that should be moved to the .rodata section. I remembered
3140 to update Builducptable as well, so it won't revert if ucptable.h is ever
3141 re-created.
3143 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3144 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3145 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3146 support is not required. This saves 1.5-2K of code, which is important in
3147 some applications.
3149 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3150 so as not to refer to the tables, even though these functions will never be
3151 called when UTF-8 support is disabled. Otherwise there are problems with a
3152 shared library.
3154 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3156 (a) It was defining its arguments as char * instead of void *.
3158 (b) It was assuming that all moves were upwards in memory; this was true
3159 a long time ago when I wrote it, but is no longer the case.
3161 The emulated memove() is provided for those environments that have neither
3162 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3163 is clearly not the case, as these two bugs were recently reported.
3165 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3166 and Detrail to create the HTML documentation, the .txt form of the man
3167 pages, and it removes trailing spaces from listed files. It also creates
3168 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3169 case, it wraps all the #defines with #ifndefs. This script should be run
3170 before "make dist".
3172 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3173 with Unicode property support.
3175 (a) For a maximizing quantifier, if the two different cases of the
3176 character were of different lengths in their UTF-8 codings (there are
3177 some cases like this - I found 11), and the matching function had to
3178 back up over a mixture of the two cases, it incorrectly assumed they
3179 were both the same length.
3181 (b) When PCRE was configured to use the heap rather than the stack for
3182 recursion during matching, it was not correctly preserving the data for
3183 the other case of a UTF-8 character when checking ahead for a match
3184 while processing a minimizing repeat. If the check also involved
3185 matching a wide character, but failed, corruption could cause an
3186 erroneous result when trying to check for a repeat of the original
3187 character.
3189 13. Some tidying changes to the testing mechanism:
3191 (a) The RunTest script now detects the internal link size and whether there
3192 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3193 values substituted by "configure". (The RunGrepTest script already did
3194 this for UTF-8.) The configure.ac script no longer substitutes the
3195 relevant variables.
3197 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3198 with length and offset values. This means that the output is different
3199 for different internal link sizes. Test 2 is skipped for link sizes
3200 other than 2 because of this, bypassing the problem. Unfortunately,
3201 there was also a test in test 3 (the locale tests) that used /B and
3202 failed for link sizes other than 2. Rather than cut the whole test out,
3203 I have added a new /Z option to pcretest that replaces the length and
3204 offset values with spaces. This is now used to make test 3 independent
3205 of link size. (Test 2 will be tidied up later.)
3207 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3208 segmentation fault instead of returning the appropriate error message.
3210 15. In multiline mode when the newline sequence was set to "any", the pattern
3211 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3212 This doesn't seem right; it now treats the CRLF combination as the line
3213 ending, and so does not match in that case. It's only a pattern such as ^$
3214 that would hit this one: something like ^ABC$ would have failed after \r
3215 and then tried again after \r\n.
3217 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3218 in an attempt to make files that differ only in their line terminators
3219 compare equal. This works on Linux.
3221 17. Under certain error circumstances pcregrep might try to free random memory
3222 as it exited. This is now fixed, thanks to valgrind.
3224 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3225 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3226 was because its rules for how to advance for /g after matching an empty
3227 string at the end of a line did not allow for this case. They now check for
3228 it specially.
3230 20. pcretest is supposed to handle patterns and data of any length, by
3231 extending its buffers when necessary. It was getting this wrong when the
3232 buffer for a data line had to be extended.
3234 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3235 CRLF as a newline sequence.
3237 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3238 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3239 I have nevertheless tidied it up.
3241 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3243 24. Added a man page for pcre-config.
3246 Version 7.0 19-Dec-06
3247 ---------------------
3249 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3250 moving to gcc 4.1.1.
3252 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3253 sys/time.h, which is documented as needed for this function. It doesn't
3254 seem to matter on Linux, but it showed up on some releases of OS X.
3256 3. It seems that there are systems where bytes whose values are greater than
3257 127 match isprint() in the "C" locale. The "C" locale should be the
3258 default when a C program starts up. In most systems, only ASCII printing
3259 characters match isprint(). This difference caused the output from pcretest
3260 to vary, making some of the tests fail. I have changed pcretest so that:
3262 (a) When it is outputting text in the compiled version of a pattern, bytes
3263 other than 32-126 are always shown as hex escapes.
3265 (b) When it is outputting text that is a matched part of a subject string,
3266 it does the same, unless a different locale has been set for the match
3267 (using the /L modifier). In this case, it uses isprint() to decide.
3269 4. Fixed a major bug that caused incorrect computation of the amount of memory
3270 required for a compiled pattern when options that changed within the
3271 pattern affected the logic of the preliminary scan that determines the
3272 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3273 that the computed length was too small. The symptoms of this bug were
3274 either the PCRE error "internal error: code overflow" from pcre_compile(),
3275 or a glibc crash with a message such as "pcretest: free(): invalid next
3276 size (fast)". Examples of patterns that provoked this bug (shown in
3277 pcretest format) are:
3279 /(?-x: )/x
3280 /(?x)(?-x: \s*#\s*)/
3281 /((?i)[\x{c0}])/8
3282 /(?i:[\x{c0}])/8
3284 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3285 is now done differently.
3287 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3288 wrapper classes; (b) implement a new function in the C++ scanner that is
3289 more efficient than the old way of doing things because it avoids levels of
3290 recursion in the regex matching; (c) add a paragraph to the documentation
3291 for the FullMatch() function.
3293 6. The escape sequence \n was being treated as whatever was defined as
3294 "newline". Not only was this contrary to the documentation, which states
3295 that \n is character 10 (hex 0A), but it also went horribly wrong when
3296 "newline" was defined as CRLF. This has been fixed.
3298 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3299 was being set to -1 for the "end of line" case (supposedly a value that no
3300 character can have). Though this value is never used (the check for end of
3301 line is "zero bytes in current character"), it caused compiler complaints.
3302 I've changed it to 0xffffffff.
3304 8. In pcre_version.c, the version string was being built by a sequence of
3305 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3306 string (as it is for production releases) called a macro with an empty
3307 argument. The C standard says the result of this is undefined. The gcc
3308 compiler treats it as an empty string (which was what was wanted) but it is
3309 reported that Visual C gives an error. The source has been hacked around to
3310 avoid this problem.
3312 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3313 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3314 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3315 of them did).
3317 10. Originally, pcretest opened its input and output without "b"; then I was
3318 told that "b" was needed in some environments, so it was added for release
3319 5.0 to both the input and output. (It makes no difference on Unix-like
3320 systems.) Later I was told that it is wrong for the input on Windows. I've
3321 now abstracted the modes into two macros, to make it easier to fiddle with
3322 them, and removed "b" from the input mode under Windows.
3324 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3326 12. Added -help and --help to pcretest as an official way of being reminded
3327 of the options.
3329 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3330 and pcrecpp.cc because they annoy compilers at high warning levels.
3332 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3334 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3335 scripts are not C programs :-) and which was not noticed because it works
3336 on Linux.
3338 16. pcretest is supposed to handle any length of pattern and data line (as one
3339 line or as a continued sequence of lines) by extending its input buffer if
3340 necessary. This feature was broken for very long pattern lines, leading to
3341 a string of junk being passed to pcre_compile() if the pattern was longer
3342 than about 50K.
3344 17. I have done a major re-factoring of the way pcre_compile() computes the
3345 amount of memory needed for a compiled pattern. Previously, there was code
3346 that made a preliminary scan of the pattern in order to do this. That was
3347 OK when PCRE was new, but as the facilities have expanded, it has become
3348 harder and harder to keep it in step with the real compile phase, and there
3349 have been a number of bugs (see for example, 4 above). I have now found a
3350 cunning way of running the real compile function in a "fake" mode that
3351 enables it to compute how much memory it would need, while actually only
3352 ever using a few hundred bytes of working memory and without too many
3353 tests of the mode. This should make future maintenance and development
3354 easier. A side effect of this work is that the limit of 200 on the nesting
3355 depth of parentheses has been removed (though this was never a serious
3356 limitation, I suspect). However, there is a downside: pcre_compile() now
3357 runs more slowly than before (30% or more, depending on the pattern). I
3358 hope this isn't a big issue. There is no effect on runtime performance.
3360 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3361 newline (only possible for the last line of a file) and it was a
3362 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3364 19. Added additional timing features to pcretest. (1) The -tm option now times
3365 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3366 separate command line item, by a number that specifies the number of
3367 repeats to use when timing. The default is 50000; this gives better
3368 precision, but takes uncomfortably long for very large patterns.
3370 20. Extended pcre_study() to be more clever in cases where a branch of a
3371 subpattern has no definite first character. For example, (a*|b*)[cd] would
3372 previously give no result from pcre_study(). Now it recognizes that the
3373 first character must be a, b, c, or d.
3375 21. There was an incorrect error "recursive call could loop indefinitely" if
3376 a subpattern (or the entire pattern) that was being tested for matching an
3377 empty string contained only one non-empty item after a nested subpattern.
3378 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3379 incorrectly, because the \d was being skipped in the check.
3381 22. The pcretest program now has a new pattern option /B and a command line
3382 option -b, which is equivalent to adding /B to every pattern. This causes
3383 it to show the compiled bytecode, without the additional information that
3384 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3385 is the same as /B/I).
3387 23. A new optimization is now able automatically to treat some sequences such
3388 as a*b as a*+b. More specifically, if something simple (such as a character
3389 or a simple class like \d) has an unlimited quantifier, and is followed by
3390 something that cannot possibly match the quantified thing, the quantifier
3391 is automatically "possessified".
3393 24. A recursive reference to a subpattern whose number was greater than 39
3394 went wrong under certain circumstances in UTF-8 mode. This bug could also
3395 have affected the operation of pcre_study().
3397 25. Realized that a little bit of performance could be had by replacing
3398 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3400 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3402 27. Possessive quantifiers such as a++ were previously implemented by turning
3403 them into atomic groups such as ($>a+). Now they have their own opcodes,
3404 which improves performance. This includes the automatically created ones
3405 from 23 above.
3407 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3408 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3409 the first matched character to be a colon. This applied both to named and
3410 numbered groups.
3412 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3414 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3415 building PCRE on Windows easier, so I have included it in the distribution.
3417 31. There is now a check in pcretest against a ridiculously large number being
3418 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3419 loop, the loop is abandoned.
3421 32. Forward references to subpatterns in conditions such as (?(2)...) where
3422 subpattern 2 is defined later cause pcre_compile() to search forwards in
3423 the pattern for the relevant set of parentheses. This search went wrong
3424 when there were unescaped parentheses in a character class, parentheses
3425 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3427 33. "Subroutine" calls and backreferences were previously restricted to
3428 referencing subpatterns earlier in the regex. This restriction has now
3429 been removed.
3431 34. Added a number of extra features that are going to be in Perl 5.10. On the
3432 whole, these are just syntactic alternatives for features that PCRE had
3433 previously implemented using the Python syntax or my own invention. The
3434 other formats are all retained for compatibility.
3436 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3437 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3438 also .NET compatible.
3440 (b) A recursion or subroutine call to a named group can now be defined as
3441 (?&name) as well as (?P>name).
3443 (c) A backreference to a named group can now be defined as \k<name> or
3444 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3445 5.10, are also .NET compatible.
3447 (d) A conditional reference to a named group can now use the syntax
3448 (?(<name>) or (?('name') as well as (?(name).
3450 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3451 groups (named and numbered) that are never evaluated inline, but can be
3452 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3453 is always false. There may be only one alternative in such a group.
3455 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3456 as the simple (?(R). The condition is true only if the most recent
3457 recursion is that of the given number or name. It does not search out
3458 through the entire recursion stack.
3460 (g) The escape \gN or \g{N} has been added, where N is a positive or
3461 negative number, specifying an absolute or relative reference.
3463 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3464 some "unreachable code" warnings.
3466 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3467 things, this adds five new scripts.
3469 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3470 There were also incompatibilities regarding the handling of \Q..\E inside
3471 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3472 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3474 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3475 matches an empty string, and forcibly breaks the loop. There were bugs in
3476 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3477 against aaaa the result was just "a" rather than "aaaa", for example. Two
3478 separate and independent bugs (that affected different cases) have been
3479 fixed.
3481 39. Refactored the code to abolish the use of different opcodes for small
3482 capturing bracket numbers. This is a tidy that I avoided doing when I
3483 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3484 The new approach is not only tidier, it makes it possible to reduce the
3485 memory needed to fix the previous bug (38).
3487 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3488 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3489 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3490 mode.
3492 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3493 report.
3495 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3496 copy construction and assignment in the C++ wrapper.
3498 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3499 couple of bugs that could have given wrong results in the "--newline=crlf"
3500 case.
3502 44. Added a number of casts and did some reorganization of signed/unsigned int
3503 variables following suggestions from Dair Grant. Also renamed the variable
3504 "this" as "item" because it is a C++ keyword.
3506 45. Arranged for dftables to add
3508 #include "pcre_internal.h"
3510 to pcre_chartables.c because without it, gcc 4.x may remove the array
3511 definition from the final binary if PCRE is built into a static library and
3512 dead code stripping is activated.
3514 46. For an unanchored pattern, if a match attempt fails at the start of a
3515 newline sequence, and the newline setting is CRLF or ANY, and the next two
3516 characters are CRLF, advance by two characters instead of one.
3519 Version 6.7 04-Jul-06
3520 ---------------------
3522 1. In order to handle tests when input lines are enormously long, pcretest has
3523 been re-factored so that it automatically extends its buffers when
3524 necessary. The code is crude, but this _is_ just a test program. The
3525 default size has been increased from 32K to 50K.
3527 2. The code in pcre_study() was using the value of the re argument before
3528 testing it for NULL. (Of course, in any sensible call of the function, it
3529 won't be NULL.)
3531 3. The memmove() emulation function in pcre_internal.h, which is used on
3532 systems that lack both memmove() and bcopy() - that is, hardly ever -
3533 was missing a "static" storage class specifier.
3535 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3536 containing an extended class (one that cannot be represented by a bitmap
3537 because it contains high-valued characters or Unicode property items, e.g.
3538 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3539 pattern, but PCRE should not loop if you do not (it no longer does).
3540 [Detail: two cases were found: (a) a repeated subpattern containing an
3541 extended class; (b) a recursive reference to a subpattern that followed a
3542 previous extended class. It wasn't skipping over the extended class
3543 correctly when UTF-8 mode was not set.]
3545 5. A negated single-character class was not being recognized as fixed-length
3546 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3547 compile error "lookbehind assertion is not fixed length".
3549 6. The RunPerlTest auxiliary script was showing an unexpected difference
3550 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3551 write a Perl script that can interpret lines of an input file either as
3552 byte characters or as UTF-8, which is what "perltest" was being required to
3553 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3554 can't do is switch easily at run time between having the "use utf8;" pragma
3555 or not. In the end, I fudged it by using the RunPerlTest script to insert
3556 "use utf8;" explicitly for the UTF-8 tests.
3558 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3559 the end of the subject string, contrary to the documentation and to what
3560 Perl does. This was true of both matching functions. Now it matches only at
3561 the start of the subject and immediately after *internal* newlines.
3563 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3564 a pointer to an int instead of a pointer to an unsigned long int. This
3565 caused problems on 64-bit systems.
3567 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3568 instance of the 'standard' template library not being so standard".
3570 10. There was no check on the number of named subpatterns nor the maximum
3571 length of a subpattern name. The product of these values is used to compute
3572 the size of the memory block for a compiled pattern. By supplying a very
3573 long subpattern name and a large number of named subpatterns, the size
3574 computation could be caused to overflow. This is now prevented by limiting
3575 the length of names to 32 characters, and the number of named subpatterns
3576 to 10,000.
3578 11. Subpatterns that are repeated with specific counts have to be replicated in
3579 the compiled pattern. The size of memory for this was computed from the
3580 length of the subpattern and the repeat count. The latter is limited to
3581 65535, but there was no limit on the former, meaning that integer overflow
3582 could in principle occur. The compiled length of a repeated subpattern is
3583 now limited to 30,000 bytes in order to prevent this.
3585 12. Added the optional facility to have named substrings with the same name.
3587 13. Added the ability to use a named substring as a condition, using the
3588 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3589 are numbers (not recommended). Forward references are permitted.
3591 14. Added forward references in named backreferences (if you see what I mean).
3593 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3594 pattern could run off the end of the subject. For example, the pattern
3595 "(?s)(.{1,5})"8 did this with the subject "ab".
3597 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3598 PCRE_CASELESS was set when matching characters that were quantified with ?
3599 or *.
3601 17. A character class other than a single negated character that had a minimum
3602 but no maximum quantifier - for example [ab]{6,} - was not handled
3603 correctly by pce_dfa_exec(). It would match only one character.
3605 18. A valid (though odd) pattern that looked like a POSIX character
3606 class but used an invalid character after [ (for example [[,abc,]]) caused
3607 pcre_compile() to give the error "Failed: internal error: code overflow" or
3608 in some cases to crash with a glibc free() error. This could even happen if
3609 the pattern terminated after [[ but there just happened to be a sequence of
3610 letters, a binary zero, and a closing ] in the memory that followed.
3612 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3613 over the years. Originally (before any Unicode support), just the bottom 8
3614 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3615 output from "man perlunicode" includes this:
3617 The regular expression compiler produces polymorphic opcodes. That
3618 is, the pattern adapts to the data and automatically switches to
3619 the Unicode character scheme when presented with Unicode data--or
3620 instead uses a traditional byte scheme when presented with byte
3621 data.
3623 Sadly, a wide octal escape does not cause a switch, and in a string with
3624 no other multibyte characters, these octal escapes are treated as before.
3625 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3626 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3627 Unicode string.
3629 I have not perpetrated such confusion in PCRE. Up till now, it took just
3630 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3631 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3632 translate to the appropriate multibyte character.
3634 29. Applied some refactoring to reduce the number of warnings from Microsoft
3635 and Borland compilers. This has included removing the fudge introduced
3636 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3637 a warning about an unused variable.
3639 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3640 characters since release 4.0, because Perl (from release 5.004) does not.
3641 [Or at least, is documented not to: some releases seem to be in conflict
3642 with the documentation.] However, when a pattern was studied with
3643 pcre_study() and all its branches started with \s, PCRE still included VT
3644 as a possible starting character. Of course, this did no harm; it just
3645 caused an unnecessary match attempt.
3647 22. Removed a now-redundant internal flag bit that recorded the fact that case
3648 dependency changed within the pattern. This was once needed for "required
3649 byte" processing, but is no longer used. This recovers a now-scarce options
3650 bit. Also moved the least significant internal flag bit to the most-
3651 significant bit of the word, which was not previously used (hangover from
3652 the days when it was an int rather than a uint) to free up another bit for
3653 the future.
3655 23. Added support for CRLF line endings as well as CR and LF. As well as the
3656 default being selectable at build time, it can now be changed at runtime
3657 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3658 specify that it is scanning data with non-default line endings.
3660 24. Changed the definition of CXXLINK to make it agree with the definition of
3661 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3663 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3664 recursions. This makes a big different to stack usage for some patterns.
3666 26. If a subpattern containing a named recursion or subroutine reference such
3667 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3668 the space required for the compiled pattern went wrong and gave too small a
3669 value. Depending on the environment, this could lead to "Failed: internal
3670 error: code overflow at offset 49" or "glibc detected double free or
3671 corruption" errors.
3673 27. Applied patches from Google (a) to support the new newline modes and (b) to
3674 advance over multibyte UTF-8 characters in GlobalReplace.
3676 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3677 difference for some implementation of PCRE in some Windows version.
3679 29. Added some extra testing facilities to pcretest:
3681 \q<number> in a data line sets the "match limit" value
3682 \Q<number> in a data line sets the "match recursion limt" value
3683 -S <number> sets the stack size, where <number> is in megabytes
3685 The -S option isn't available for Windows.
3688 Version 6.6 06-Feb-06
3689 ---------------------
3691 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3692 in pcreposix.h. I have copied the definition from pcre.h.
3694 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3695 because pcre.h is no longer a built file.
3697 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3698 not normally included in the compiled code.
3701 Version 6.5 01-Feb-06
3702 ---------------------
3704 1. When using the partial match feature with pcre_dfa_exec(), it was not
3705 anchoring the second and subsequent partial matches at the new starting
3706 point. This could lead to incorrect results. For example, with the pattern
3707 /1234/, partially matching against "123" and then "a4" gave a match.
3709 2. Changes to pcregrep:
3711 (a) All non-match returns from pcre_exec() were being treated as failures
3712 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3713 error message is output. Some extra information is given for the
3715 probably the only errors that are likely to be caused by users (by
3716 specifying a regex that has nested indefinite repeats, for instance).
3717 If there are more than 20 of these errors, pcregrep is abandoned.
3719 (b) A binary zero was treated as data while matching, but terminated the
3720 output line if it was written out. This has been fixed: binary zeroes
3721 are now no different to any other data bytes.
3723 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3724 used to set a locale for matching. The --locale=xxxx long option has
3725 been added (no short equivalent) to specify a locale explicitly on the
3726 pcregrep command, overriding the environment variables.
3728 (d) When -B was used with -n, some line numbers in the output were one less
3729 than they should have been.
3731 (e) Added the -o (--only-matching) option.
3733 (f) If -A or -C was used with -c (count only), some lines of context were
3734 accidentally printed for the final match.
3736 (g) Added the -H (--with-filename) option.
3738 (h) The combination of options -rh failed to suppress file names for files
3739 that were found from directory arguments.
3741 (i) Added the -D (--devices) and -d (--directories) options.
3743 (j) Added the -F (--fixed-strings) option.
3745 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3747 (l) Added the --colo(u)r option.
3749 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3750 is not present by default.
3752 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3753 items such as (?R) or (?1), when the recursion could match a number of
3754 alternatives. If it matched one of the alternatives, but subsequently,
3755 outside the recursion, there was a failure, the code tried to back up into
3756 the recursion. However, because of the way PCRE is implemented, this is not
3757 possible, and the result was an incorrect result from the match.
3759 In order to prevent this happening, the specification of recursion has
3760 been changed so that all such subpatterns are automatically treated as
3761 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3763 4. I had overlooked the fact that, in some locales, there are characters for
3764 which isalpha() is true but neither isupper() nor islower() are true. In
3765 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3766 and ordfeminine) are like this. This affected the treatment of \w and \W
3767 when they appeared in character classes, but not when they appeared outside
3768 a character class. The bit map for "word" characters is now created
3769 separately from the results of isalnum() instead of just taking it from the
3770 upper, lower, and digit maps. (Plus the underscore character, of course.)
3772 5. The above bug also affected the handling of POSIX character classes such as
3773 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3774 permanent tables. Instead, the bit maps for such a class were previously
3775 created as the appropriate unions of the upper, lower, and digit bitmaps.
3776 Now they are created by subtraction from the [[:word:]] class, which has
3777 its own bitmap.
3779 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3780 It is created by subtracting the vertical space characters (\x09, \x0a,
3781 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3782 subtraction was done in the overall bitmap for a character class, meaning
3783 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3784 be recognized. This bug has been fixed.
3786 7. Patches from the folks at Google:
3788 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3789 real life, but is still worth protecting against".
3791 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3792 regular expressions".
3794 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3795 have it.
3797 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3798 "configure" and the latter not, in order to fix a problem somebody had
3799 with compiling the Arg class on HP-UX.
3801 (e) Improve the error-handling of the C++ wrapper a little bit.
3803 (f) New tests for checking recursion limiting.
3805 8. The pcre_memmove() function, which is used only if the environment does not
3806 have a standard memmove() function (and is therefore rarely compiled),
3807 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3808 returning a result (though PCRE never actually uses the result).
3810 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3811 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3812 returned instead of calling malloc() with an overflowing number that would
3813 most likely cause subsequent chaos.
3815 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3817 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3818 with this option is matched, the nmatch and pmatch options of regexec() are
3819 ignored.
3821 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3822 provided in case anyone wants to the the POSIX interface with UTF-8
3823 strings.
3825 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3826 C++ linking (needed for some HP-UX environments).
3828 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3829 (unused parameter) and in the pcre_printint() function (omitted "default"
3830 switch label when the default is to do nothing).
3832 15. Added some code to make it possible, when PCRE is compiled as a C++
3833 library, to replace subject pointers for pcre_exec() with a smart pointer
3834 class, thus making it possible to process discontinuous strings.
3836 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3837 much the same function. They were added by different people who were trying
3838 to make PCRE easy to compile on non-Unix systems. It has been suggested
3839 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3840 for compiling on Windows systems. I have therefore replaced it with
3841 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3842 defaults to "extern" for C or "extern C" for C++, which works fine on
3843 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3844 SCOPE with something explicit in config.h. In addition:
3846 (a) pcreposix.h still had just "extern" instead of either of these macros;
3847 I have replaced it with PCRE_DATA_SCOPE.
3849 (b) Functions such as _pcre_xclass(), which are internal to the library,
3850 but external in the C sense, all had PCRE_EXPORT in their definitions.
3851 This is apparently wrong for the Windows case, so I have removed it.
3852 (It makes no difference on Unix-like systems.)
3854 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3855 of recursive calls to match(). This is different to MATCH_LIMIT because
3856 that limits the total number of calls to match(), not all of which increase
3857 the depth of recursion. Limiting the recursion depth limits the amount of
3858 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3859 when PCRE is compiled, and changed at run time. A patch from Google adds
3860 this functionality to the C++ interface.
3862 18. Changes to the handling of Unicode character properties:
3864 (a) Updated the table to Unicode 4.1.0.
3866 (b) Recognize characters that are not in the table as "Cn" (undefined).
3868 (c) I revised the way the table is implemented to a much improved format
3869 which includes recognition of ranges. It now supports the ranges that
3870 are defined in UnicodeData.txt, and it also amalgamates other
3871 characters into ranges. This has reduced the number of entries in the
3872 table from around 16,000 to around 3,000, thus reducing its size
3873 considerably. I realized I did not need to use a tree structure after
3874 all - a binary chop search is just as efficient. Having reduced the
3875 number of entries, I extended their size from 6 bytes to 8 bytes to
3876 allow for more data.
3878 (d) Added support for Unicode script names via properties such as \p{Han}.
3880 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3881 matching that character.
3883 20. When matching a repeated Unicode property with a minimum greater than zero,
3884 (for example \pL{2,}), PCRE could look past the end of the subject if it
3885 reached it while seeking the minimum number of characters. This could
3886 happen only if some of the characters were more than one byte long, because
3887 there is a check for at least the minimum number of bytes.
3889 21. Refactored the implementation of \p and \P so as to be more general, to
3890 allow for more different types of property in future. This has changed the
3891 compiled form incompatibly. Anybody with saved compiled patterns that use
3892 \p or \P will have to recompile them.
3894 22. Added "Any" and "L&" to the supported property types.
3896 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3897 but give a compile time error if the value is greater than 0xff.
3899 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3900 accidentally not being installed or uninstalled.
3902 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3903 made were to insert the current release number. This seemed silly, because
3904 it made things harder for people building PCRE on systems that don't run
3905 "configure". I have turned pcre.h into a distributed file, no longer built
3906 by "configure", with the version identification directly included. There is
3907 no longer a pcre.h.in file.
3909 However, this change necessitated a change to the pcre-config script as
3910 well. It is built from pcre-config.in, and one of the substitutions was the
3911 release number. I have updated configure.ac so that ./configure now finds
3912 the release number by grepping pcre.h.
3914 26. Added the ability to run the tests under valgrind.
3917 Version 6.4 05-Sep-05
3918 ---------------------
3920 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3921 "--" to be printed when multiple files were scanned, even when none of the
3922 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3923 consider it to be a bug, and have restored the previous behaviour.
3925 2. A couple of code tidies to get rid of compiler warnings.
3927 3. The pcretest program used to cheat by referring to symbols in the library
3928 whose names begin with _pcre_. These are internal symbols that are not
3929 really supposed to be visible externally, and in some environments it is
3930 possible to suppress them. The cheating is now confined to including
3931 certain files from the library's source, which is a bit cleaner.
3933 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3934 file's purpose clearer.
3936 5. Reorganized pcre_ucp_findchar().
3939 Version 6.3 15-Aug-05
3940 ---------------------
3942 1. The file libpcre.pc.in did not have general read permission in the tarball.
3944 2. There were some problems when building without C++ support:
3946 (a) If C++ support was not built, "make install" and "make test" still
3947 tried to test it.
3949 (b) There were problems when the value of CXX was explicitly set. Some
3950 changes have been made to try to fix these, and ...
3952 (c) --disable-cpp can now be used to explicitly disable C++ support.
3954 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3955 backslash in a target when C++ was disabled. This confuses some
3956 versions of "make", apparently. Using an intermediate variable solves
3957 this. (Same for CPP_LOBJ.)
3959 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3960 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3961 necessary on certain architectures.
3963 4. Added a setting of -export-symbols-regex to the link command to remove
3964 those symbols that are exported in the C sense, but actually are local
3965 within the library, and not documented. Their names all begin with
3966 "_pcre_". This is not a perfect job, because (a) we have to except some
3967 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3968 available (and never for static libraries). I have made a note to try to
3969 find a way round (a) in the future.
3972 Version 6.2 01-Aug-05
3973 ---------------------
3975 1. There was no test for integer overflow of quantifier values. A construction
3976 such as {1111111111111111} would give undefined results. What is worse, if
3977 a minimum quantifier for a parenthesized subpattern overflowed and became
3978 negative, the calculation of the memory size went wrong. This could have
3979 led to memory overwriting.
3981 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3983 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3984 operating environments where this matters.
3986 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3987 PCRE options from within the C++ wrapper.
3989 5. Named capturing subpatterns were not being correctly counted when a pattern
3990 was compiled. This caused two problems: (a) If there were more than 100
3991 such subpatterns, the calculation of the memory needed for the whole
3992 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3993 back references of the form \12, where the number was greater than 9, were
3994 not recognized as back references, even though there were sufficient
3995 previous subpatterns.
3997 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3998 versions of gcc, e.g. 2.95.4.
4001 Version 6.1 21-Jun-05
4002 ---------------------
4004 1. There was one reference to the variable "posix" in pcretest.c that was not
4005 surrounded by "#if !defined NOPOSIX".
4007 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4008 the cross-check on the old pcre_info() function, for the benefit of the
4009 cut-down version of PCRE that is currently imported into Exim.
4011 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4012 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4013 patterns that start (?i) but I don't think that matters. The use of (?i) is
4014 just an example; this all applies to the other options as well.
4016 4. Since libtool seems to echo the compile commands it is issuing, the output
4017 from "make" can be reduced a bit by putting "@" in front of each libtool
4018 compile command.
4020 5. Patch from the folks at Google for configure.in to be a bit more thorough
4021 in checking for a suitable C++ installation before trying to compile the
4022 C++ stuff. This should fix a reported problem when a compiler was present,
4023 but no suitable headers.
4025 6. The man pages all had just "PCRE" as their title. I have changed them to
4026 be the relevant file name. I have also arranged that these names are
4027 retained in the file doc/pcre.txt, which is a concatenation in text format
4028 of all the man pages except the little individual ones for each function.
4030 7. The NON-UNIX-USE file had not been updated for the different set of source
4031 files that come with release 6. I also added a few comments about the C++
4032 wrapper.
4035 Version 6.0 07-Jun-05
4036 ---------------------
4038 1. Some minor internal re-organization to help with my DFA experiments.
4040 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4041 didn't matter for the library itself when fully configured, but did matter
4042 when compiling without UCP support, or within Exim, where the ucp files are
4043 not imported.
4045 3. Refactoring of the library code to split up the various functions into
4046 different source modules. The addition of the new DFA matching code (see
4047 below) to a single monolithic source would have made it really too
4048 unwieldy, quite apart from causing all the code to be include in a
4049 statically linked application, when only some functions are used. This is
4050 relevant even without the DFA addition now that patterns can be compiled in
4051 one application and matched in another.
4053 The downside of splitting up is that there have to be some external
4054 functions and data tables that are used internally in different modules of
4055 the library but which are not part of the API. These have all had their
4056 names changed to start with "_pcre_" so that they are unlikely to clash
4057 with other external names.
4059 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4060 a different (DFA) algorithm. Although it is slower than the original
4061 function, it does have some advantages for certain types of matching
4062 problem.
4064 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4065 including restarting after a partial match.
4067 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4068 defined when compiling for Windows was sent to me. I have put it into the
4069 code, though I have no means of testing or verifying it.
4071 7. Added the pcre_refcount() auxiliary function.
4073 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4074 match before or at the first newline in the subject string. In pcretest,
4075 the /f option on a pattern can be used to set this.
4077 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4078 would behave wrongly. This has been present in PCRE since release 4.0.
4080 10. A number of changes to the pcregrep command:
4082 (a) Refactored how -x works; insert ^(...)$ instead of setting
4083 PCRE_ANCHORED and checking the length, in preparation for adding
4084 something similar for -w.
4086 (b) Added the -w (match as a word) option.
4088 (c) Refactored the way lines are read and buffered so as to have more
4089 than one at a time available.
4091 (d) Implemented a pcregrep test script.
4093 (e) Added the -M (multiline match) option. This allows patterns to match
4094 over several lines of the subject. The buffering ensures that at least
4095 8K, or the rest of the document (whichever is the shorter) is available
4096 for matching (and similarly the previous 8K for lookbehind assertions).
4098 (f) Changed the --help output so that it now says
4100 -w, --word-regex(p)
4102 instead of two lines, one with "regex" and the other with "regexp"
4103 because that confused at least one person since the short forms are the
4104 same. (This required a bit of code, as the output is generated
4105 automatically from a table. It wasn't just a text change.)
4107 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4108 option but starts with a hyphen. Could be a pattern or a path name
4109 starting with a hyphen, for instance.
4111 (h) "-" can be given as a file name to represent stdin.
4113 (i) When file names are being printed, "(standard input)" is used for
4114 the standard input, for compatibility with GNU grep. Previously
4115 "<stdin>" was used.
4117 (j) The option --label=xxx can be used to supply a name to be used for
4118 stdin when file names are being printed. There is no short form.
4120 (k) Re-factored the options decoding logic because we are going to add
4121 two more options that take data. Such options can now be given in four
4122 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4124 (l) Added the -A, -B, and -C options for requesting that lines of context
4125 around matches be printed.
4127 (m) Added the -L option to print the names of files that do not contain
4128 any matching lines, that is, the complement of -l.
4130 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4131 continue to scan other files.
4133 (o) The -s option was incorrectly implemented. For compatibility with other
4134 greps, it now suppresses the error message for a non-existent or non-
4135 accessible file (but not the return code). There is a new option called
4136 -q that suppresses the output of matching lines, which was what -s was
4137 previously doing.
4139 (p) Added --include and --exclude options to specify files for inclusion
4140 and exclusion when recursing.
4142 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4143 Hopefully, it now does.
4145 12. Missing cast in pcre_study().
4147 13. Added an "uninstall" target to the makefile.
4149 14. Replaced "extern" in the function prototypes in Makefile.in with
4150 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4151 world, but is set differently for Windows.
4153 15. Added a second compiling function called pcre_compile2(). The only
4154 difference is that it has an extra argument, which is a pointer to an
4155 integer error code. When there is a compile-time failure, this is set
4156 non-zero, in addition to the error test pointer being set to point to an
4157 error message. The new argument may be NULL if no error number is required
4158 (but then you may as well call pcre_compile(), which is now just a
4159 wrapper). This facility is provided because some applications need a
4160 numeric error indication, but it has also enabled me to tidy up the way
4161 compile-time errors are handled in the POSIX wrapper.
4163 16. Added VPATH=.libs to the makefile; this should help when building with one
4164 prefix path and installing with another. (Or so I'm told by someone who
4165 knows more about this stuff than I do.)
4167 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4168 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4169 match everything, including newlines. This is not POSIX-compatible, but
4170 somebody wanted the feature. From pcretest it can be activated by using
4171 both the P and the s flags.
4173 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4175 19. libpcre.pc was being incorrectly installed as executable.
4177 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4178 it now also looks for '\r' so that it will work unmodified on Windows.
4180 21. Added Google's contributed C++ wrapper to the distribution.
4182 22. Added some untidy missing memory free() calls in pcretest, to keep
4183 Electric Fence happy when testing.
4187 Version 5.0 13-Sep-04
4188 ---------------------
4190 1. Internal change: literal characters are no longer packed up into items
4191 containing multiple characters in a single byte-string. Each character
4192 is now matched using a separate opcode. However, there may be more than one
4193 byte in the character in UTF-8 mode.
4195 2. The pcre_callout_block structure has two new fields: pattern_position and
4196 next_item_length. These contain the offset in the pattern to the next match
4197 item, and its length, respectively.
4199 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4200 insertion of callouts before each pattern item. Added the /C option to
4201 pcretest to make use of this.
4203 4. On the advice of a Windows user, the lines
4205 #if defined(_WIN32) || defined(WIN32)
4206 _setmode( _fileno( stdout ), 0x8000 );
4207 #endif /* defined(_WIN32) || defined(WIN32) */
4209 have been added to the source of pcretest. This apparently does useful
4210 magic in relation to line terminators.
4212 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4213 for the benefit of those environments where the "b" makes a difference.
4215 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4216 to know about it. I have put a hack into configure.in that adds in code
4217 to set GCC=yes if CC=icc. This seems to end up at a point in the
4218 generated configure script that is early enough to affect the setting of
4219 compiler options, which is what is needed, but I have no means of testing
4220 whether it really works. (The user who reported this had patched the
4221 generated configure script, which of course I cannot do.)
4223 LATER: After change 22 below (new libtool files), the configure script
4224 seems to know about icc (and also ecc). Therefore, I have commented out
4225 this hack in configure.in.
4227 7. Added support for pkg-config (2 patches were sent in).
4229 8. Negated POSIX character classes that used a combination of internal tables
4230 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4231 [[:^ascii]]. Typically, they would match almost any characters. The other
4232 POSIX classes were not broken in this way.
4234 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4235 to find the match, as PCRE was deluded into thinking that the match had to
4236 start at the start point or following a newline. The same bug applied to
4237 patterns with negative forward assertions or any backward assertions
4238 preceding ".*" at the start, unless the pattern required a fixed first
4239 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4241 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4242 starting at the last subject character, bytes beyond the end of the subject
4243 string were read.
4245 11. Renamed the variable "class" as "classbits" to make life easier for C++
4246 users. (Previously there was a macro definition, but it apparently wasn't
4247 enough.)
4249 12. Added the new field "tables" to the extra data so that tables can be passed
4250 in at exec time, or the internal tables can be re-selected. This allows
4251 a compiled regex to be saved and re-used at a later time by a different
4252 program that might have everything at different addresses.
4254 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4255 -R library as well as a -L library.
4257 14. The debugging options of pcretest (-d on the command line or D on a
4258 pattern) showed incorrect output for anything following an extended class
4259 that contained multibyte characters and which was followed by a quantifier.
4261 15. Added optional support for general category Unicode character properties
4262 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4263 support. It adds about 90K to the size of the library. The meanings of the
4264 inbuilt class escapes such as \d and \s have NOT been changed.
4266 16. Updated pcredemo.c to include calls to free() to release the memory for the
4267 compiled pattern.
4269 17. The generated file chartables.c was being created in the source directory
4270 instead of in the building directory. This caused the build to fail if the
4271 source directory was different from the building directory, and was
4272 read-only.
4274 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4275 file. No doubt somebody will tell me if they don't make sense... Also added
4276 Dan Mooney's comments about building on OpenVMS.
4278 19. Added support for partial matching via the PCRE_PARTIAL option for
4279 pcre_exec() and the \P data escape in pcretest.
4281 20. Extended pcretest with 3 new pattern features:
4283 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4284 write the compiled pattern to the file whose name is "rest-of-line".
4285 This is a straight binary dump of the data, with the saved pointer to
4286 the character tables forced to be NULL. The study data, if any, is
4287 written too. After writing, pcretest reads a new pattern.
4289 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4290 compiled pattern from the given file. There must not be any
4291 occurrences of "<" in the file name (pretty unlikely); if there are,
4292 pcretest will instead treat the initial "<" as a pattern delimiter.
4293 After reading in the pattern, pcretest goes on to read data lines as
4294 usual.
4296 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4297 and 16-bit fields in a compiled pattern, to simulate a pattern that
4298 was compiled on a host of opposite endianness.
4300 21. The pcre-exec() function can now cope with patterns that were compiled on
4301 hosts of opposite endianness, with this restriction:
4303 As for any compiled expression that is saved and used later, the tables
4304 pointer field cannot be preserved; the extra_data field in the arguments
4305 to pcre_exec() should be used to pass in a tables address if a value
4306 other than the default internal tables were used at compile time.
4308 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4309 now diagnosed as an error. Previously, most of the time, a negative number
4310 would have been treated as zero, but if in addition "ovector" was passed as
4311 NULL, a crash could occur.
4313 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4314 new versions from the libtool 1.5 distribution (the last one is a copy of
4315 a file called libtool.m4). This seems to have fixed the need to patch
4316 "configure" to support Darwin 1.3 (which I used to do). However, I still
4317 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4318 workstation).
4320 24. Changed the PCRE licence to be the more standard "BSD" licence.
4323 Version 4.5 01-Dec-03
4324 ---------------------
4326 1. There has been some re-arrangement of the code for the match() function so
4327 that it can be compiled in a version that does not call itself recursively.
4328 Instead, it keeps those local variables that need separate instances for
4329 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4330 needs to "recurse". Keeping track of where control must go is done by means
4331 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4332 hide most of the details from the main code, and operates only if
4333 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4334 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4335 operating.
4337 To make it easier for callers to provide specially tailored get/free
4338 functions for this usage, two new functions, pcre_stack_malloc, and
4339 pcre_stack_free, are used. They are always called in strict stacking order,
4340 and the size of block requested is always the same.
4342 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4343 PCRE has been compiled to use the stack or the heap for recursion. The
4344 -C option of pcretest uses this to show which version is compiled.
4346 A new data escape \S, is added to pcretest; it causes the amounts of store
4347 obtained and freed by both kinds of malloc/free at match time to be added
4348 to the output.
4350 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4351 what's available on my current Linux desktop machine.
4353 3. When matching a UTF-8 string, the test for a valid string at the start has
4354 been extended. If start_offset is not zero, PCRE now checks that it points
4355 to a byte that is the start of a UTF-8 character. If not, it returns
4356 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4357 this is necessary because there may be backward assertions in the pattern.
4358 When matching the same subject several times, it may save resources to use
4359 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4361 4. The code for checking the validity of UTF-8 strings has been tightened so
4362 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4363 containing "overlong sequences".
4365 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4366 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4367 should have been "&", but it just so happened that all the cases this let
4368 through by mistake were picked up later in the function.
4370 6. I had used a variable called "isblank" - this is a C99 function, causing
4371 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4373 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4374 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4375 so the version is shown; (c) stop "make test" echoing "./RunTest".
4377 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4379 9. The prototype for memmove() for systems that don't have it was using
4380 size_t, but the inclusion of the header that defines size_t was later. I've
4381 moved the #includes for the C headers earlier to avoid this.
4383 10. Added some adjustments to the code to make it easier to compiler on certain
4384 special systems:
4386 (a) Some "const" qualifiers were missing.
4387 (b) Added the macro EXPORT before all exported functions; by default this
4388 is defined to be empty.
4389 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4390 that it reads its output file name as an argument instead of writing
4391 to the standard output and assuming this can be redirected.
4393 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4394 class containing characters with values greater than 255, PCRE compilation
4395 went into a loop.
4397 12. A recursive reference to a subpattern that was within another subpattern
4398 that had a minimum quantifier of zero caused PCRE to crash. For example,
4399 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4400 recursion. If the recursively-called subpattern itself had a zero repeat,
4401 that was OK.
4403 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4404 buffer into which it was copied (for escape processing) was still set at
4405 1024, so long lines caused crashes.
4407 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4408 "internal error: code overflow...". This applied to any character class
4409 that was followed by a possessive quantifier.
4411 15. Modified the Makefile to add libpcre.la as a prerequisite for
4412 libpcreposix.la because I was told this is needed for a parallel build to
4413 work.
4415 16. If a pattern that contained .* following optional items at the start was
4416 studied, the wrong optimizing data was generated, leading to matching
4417 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4418 matching string must start with a or b or c. The correct conclusion for
4419 this pattern is that a match can start with any character.
4422 Version 4.4 13-Aug-03
4423 ---------------------
4425 1. In UTF-8 mode, a character class containing characters with values between
4426 127 and 255 was not handled correctly if the compiled pattern was studied.
4427 In fixing this, I have also improved the studying algorithm for such
4428 classes (slightly).
4430 2. Three internal functions had redundant arguments passed to them. Removal
4431 might give a very teeny performance improvement.
4433 3. Documentation bug: the value of the capture_top field in a callout is *one
4434 more than* the number of the hightest numbered captured substring.
4436 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4437 in incorrectly linking with a previously installed version. They now link
4438 explicitly with libpcre.la.
4440 5. configure.in no longer needs to recognize Cygwin specially.
4442 6. A problem in pcre.in for Windows platforms is fixed.
4444 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4445 pcretest, it used to include the size of the study block as part of its
4446 output. Unfortunately, the structure contains a field that has a different
4447 size on different hardware architectures. This meant that the tests that
4448 showed this size failed. As the block is currently always of a fixed size,
4449 this information isn't actually particularly useful in pcretest output, so
4450 I have just removed it.
4452 8. Three pre-processor statements accidentally did not start in column 1.
4453 Sadly, there are *still* compilers around that complain, even though
4454 standard C has not required this for well over a decade. Sigh.
4456 9. In pcretest, the code for checking callouts passed small integers in the
4457 callout_data field, which is a void * field. However, some picky compilers
4458 complained about the casts involved for this on 64-bit systems. Now
4459 pcretest passes the address of the small integer instead, which should get
4460 rid of the warnings.
4462 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4463 both compile and run time, and gives an error if an invalid UTF-8 sequence
4464 is found. There is a option for disabling this check in cases where the
4465 string is known to be correct and/or the maximum performance is wanted.
4467 11. In response to a bug report, I changed one line in Makefile.in from
4469 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4470 to
4471 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4473 to look similar to other lines, but I have no way of telling whether this
4474 is the right thing to do, as I do not use Windows. No doubt I'll get told
4475 if it's wrong...
4478 Version 4.3 21-May-03
4479 ---------------------
4481 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4482 Makefile.
4484 2. Some refactoring to improve the quality of the code:
4486 (i) The utf8_table... variables are now declared "const".
4488 (ii) The code for \cx, which used the "case flipping" table to upper case
4489 lower case letters, now just substracts 32. This is ASCII-specific,
4490 but the whole concept of \cx is ASCII-specific, so it seems
4491 reasonable.
4493 (iii) PCRE was using its character types table to recognize decimal and
4494 hexadecimal digits in the pattern. This is silly, because it handles
4495 only 0-9, a-f, and A-F, but the character types table is locale-
4496 specific, which means strange things might happen. A private
4497 table is now used for this - though it costs 256 bytes, a table is
4498 much faster than multiple explicit tests. Of course, the standard
4499 character types table is still used for matching digits in subject
4500 strings against \d.
4502 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4503 ending in _t are). So I've renamed it as ESC_tee.
4505 3. The first argument for regexec() in the POSIX wrapper should have been
4506 defined as "const".
4508 4. Changed pcretest to use malloc() for its buffers so that they can be
4509 Electric Fenced for debugging.
4511 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4512 to read one or more bytes before the start of the subject string. Often this
4513 had no effect on PCRE's behaviour, but in some circumstances it could
4514 provoke a segmentation fault.
4516 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4517 to try to read one or more bytes before the start of the subject string.
4519 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4520 UTF-8 support could misbehave in various ways if the subject string
4521 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4522 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4523 back over UTF-8 characters.)
4526 Version 4.2 14-Apr-03
4527 ---------------------
4529 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4531 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4532 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4533 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4534 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4535 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4537 Note: automatic setting of the BUILD variables is not yet working
4538 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4539 compile-time but not at link-time
4540 [LINK]: use for linking executables only
4541 make different versions for Windows and non-Windows
4542 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4543 libraries
4544 [LINK_FOR_BUILD]: new variable
4545 [OBJEXT]: use throughout
4546 [EXEEXT]: use throughout
4547 <winshared>: new target
4548 <wininstall>: new target
4549 <dftables.o>: use native compiler
4550 <dftables>: use native linker
4551 <install>: handle Windows platform correctly
4552 <clean>: ditto
4553 <check>: ditto
4554 copy DLL to top builddir before testing
4556 As part of these changes, -no-undefined was removed again. This was reported
4557 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4558 in any case.
4560 3. Some tidies to get rid of compiler warnings:
4562 . In the match_data structure, match_limit was an unsigned long int, whereas
4563 match_call_count was an int. I've made them both unsigned long ints.
4565 . In pcretest the fact that a const uschar * doesn't automatically cast to
4566 a void * provoked a warning.
4568 . Turning on some more compiler warnings threw up some "shadow" variables
4569 and a few more missing casts.
4571 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4572 option, a class that contained a single character with a value between 128
4573 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4575 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4576 option, a class that contained several characters, but with at least one
4577 whose value was between 128 and 255 caused PCRE to crash.
4580 Version 4.1 12-Mar-03
4581 ---------------------
4583 1. Compiling with gcc -pedantic found a couple of places where casts were
4584 needed, and a string in dftables.c that was longer than standard compilers are
4585 required to support.
4587 2. Compiling with Sun's compiler found a few more places where the code could
4588 be tidied up in order to avoid warnings.
4590 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4591 first of these names is deprecated in the latest Autoconf in favour of the name
4592 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4593 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4594 analogy I have changed it to CFLAGS_FOR_BUILD.
4596 4. Added -no-undefined to the linking command in the Makefile, because this is
4597 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4598 linking step for the pcreposix library.
4600 5. PCRE was failing to diagnose the case of two named groups with the same
4601 name.
4603 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4604 literal character that is needed in the subject for a match, and scans along to
4605 ensure that it is present before embarking on the full matching process. This
4606 saves time in cases of nested unlimited repeats that are never going to match.
4607 Problem: the scan can take a lot of time if the subject is very long (e.g.
4608 megabytes), thus penalizing straightforward matches. It is now done only if the
4609 amount of subject to be scanned is less than 1000 bytes.
4611 7. A lesser problem with the same optimization is that it was recording the
4612 first character of an anchored pattern as "needed", thus provoking a search
4613 right along the subject, even when the first match of the pattern was going to
4614 fail. The "needed" character is now not set for anchored patterns, unless it
4615 follows something in the pattern that is of non-fixed length. Thus, it still
4616 fulfils its original purpose of finding quick non-matches in cases of nested
4617 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4620 Version 4.0 17-Feb-03
4621 ---------------------
4623 1. If a comment in an extended regex that started immediately after a meta-item
4624 extended to the end of string, PCRE compiled incorrect data. This could lead to
4625 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4627 2. Moved to autoconf 2.53 and libtool 1.4.2.
4629 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4630 the special perltest8 script is no longer needed - all the tests can be run
4631 from a single perltest script.
4633 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4634 by \s. It has now been removed in PCRE. This means it isn't recognized as
4635 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4636 class [:space:] *does* include VT, thereby creating a mess.
4638 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4639 space and tab.
4641 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4642 its new features into the main test script, reducing the number of scripts.
4644 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4645 were backward compatible, and made the (?i) apply to the whole pattern, as if
4646 /i were given. Now it behaves more logically, and applies the option setting
4647 only to what follows. PCRE has been changed to follow suit. However, if it
4648 finds options settings right at the start of the pattern, it extracts them into
4649 the global options, as before. Thus, they show up in the info data.
4651 8. Added support for the \Q...\E escape sequence. Characters in between are
4652 treated as literals. This is slightly different from Perl in that $ and @ are
4653 also handled as literals inside the quotes. In Perl, they will cause variable
4654 interpolation. Note the following examples:
4656 Pattern PCRE matches Perl matches
4658 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4659 \Qabc\$xyz\E abc\$xyz abc\$xyz
4660 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4662 For compatibility with Perl, \Q...\E sequences are recognized inside character
4663 classes as well as outside them.
4665 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4666 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4667 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4668 signed/unsigned warnings.
4670 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4671 option for pcretest, so I've replaced it by a simple function that does just
4672 that job.
4674 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4675 "pcregrep -".
4677 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4678 Java package. This provides some syntactic sugar for simple cases of what my
4679 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4680 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4681 item, you can use this simplified notation. Note that only makes sense with
4682 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4683 greediness, whatever the setting of the PCRE_UNGREEDY option.
4685 13. A change of greediness default within a pattern was not taking effect at
4686 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4687 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4688 was abstracted outside.
4690 14. PCRE now supports the \G assertion. It is true when the current matching
4691 position is at the start point of the match. This differs from \A when the
4692 starting offset is non-zero. Used with the /g option of pcretest (or similar
4693 code), it works in the same way as it does for Perl's /g option. If all
4694 alternatives of a regex begin with \G, the expression is anchored to the start
4695 match position, and the "anchored" flag is set in the compiled expression.
4697 15. Some bugs concerning the handling of certain option changes within patterns
4698 have been fixed. These applied to options other than (?ims). For example,
4699 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4700 been the other way round. Some of this was related to change 7 above.
4702 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4703 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4704 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4705 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4707 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4708 mode. Unlike ".", it always matches newline, whatever the setting of
4709 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4710 assertions. Perl allows it, but it doesn't (in general) work because it can't
4711 calculate the length of the lookbehind. At least, that's the case for Perl
4712 5.8.0 - I've been told they are going to document that it doesn't work in
4713 future.
4715 18. Added an error diagnosis for escapes that PCRE does not support: these are
4716 \L, \l, \N, \P, \p, \U, \u, and \X.
4718 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4719 reading past the end of the pattern in cases such as /[abcd/.
4721 20. PCRE was getting more memory than necessary for patterns with classes that
4722 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4724 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4725 compiling PCRE for use with Virtual Pascal.
4727 22. Small fix to the Makefile to make it work properly if the build is done
4728 outside the source tree.
4730 23. Added a new extension: a condition to go with recursion. If a conditional
4731 subpattern starts with (?(R) the "true" branch is used if recursion has
4732 happened, whereas the "false" branch is used only at the top level.
4734 24. When there was a very long string of literal characters (over 255 bytes
4735 without UTF support, over 250 bytes with UTF support), the computation of how
4736 much memory was required could be incorrect, leading to segfaults or other
4737 strange effects.
4739 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4740 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4741 there was a subsequent back reference to those brackets. This meant that, for
4742 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4743 possible to check for precisely this case. All we can do is abandon the
4744 optimization if .* occurs inside capturing brackets when there are any back
4745 references whatsoever. (See below for a better fix that came later.)
4747 26. The handling of the optimization for finding the first character of a
4748 non-anchored pattern, and for finding a character that is required later in the
4749 match were failing in some cases. This didn't break the matching; it just
4750 failed to optimize when it could. The way this is done has been re-implemented.
4752 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4754 28. Added a new feature that provides some of the functionality that Perl
4755 provides with (?{...}). The facility is termed a "callout". The way it is done
4756 in PCRE is for the caller to provide an optional function, by setting
4757 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4758 global variable. By default it is unset, which disables all calling out. To get
4759 the function called, the regex must include (?C) at appropriate points. This
4760 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4761 This provides a means of identifying different callout points. When PCRE
4762 reaches such a point in the regex, if pcre_callout has been set, the external
4763 function is called. It is provided with data in a structure called
4764 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4765 matching continues; if it returns a non-zero value, the match at the current
4766 point fails. However, backtracking will occur if possible. [This was changed
4767 later and other features added - see item 49 below.]
4769 29. pcretest is upgraded to test the callout functionality. It provides a
4770 callout function that displays information. By default, it shows the start of
4771 the match and the current position in the text. There are some new data escapes
4772 to vary what happens:
4774 \C+ in addition, show current contents of captured substrings
4775 \C- do not supply a callout function
4776 \C!n return 1 when callout number n is reached
4777 \C!n!m return 1 when callout number n is reached for the mth time
4779 30. If pcregrep was called with the -l option and just a single file name, it
4780 output "<stdin>" if a match was found, instead of the file name.
4782 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4783 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4784 pcre_exec(). This saves a malloc/free per call. The default value of
4785 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4786 when configuring.
4788 32. The default maximum size of a compiled pattern is 64K. There have been a
4789 few cases of people hitting this limit. The code now uses macros to handle the
4790 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4791 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4792 configuring. Tests 2 and 5 work only with 2-byte links because they output
4793 debugging information about compiled patterns.
4795 33. Internal code re-arrangements:
4797 (a) Moved the debugging function for printing out a compiled regex into
4798 its own source file (printint.c) and used #include to pull it into
4799 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4800 separate copies.
4802 (b) Defined the list of op-code names for debugging as a macro in
4803 internal.h so that it is next to the definition of the opcodes.
4805 (c) Defined a table of op-code lengths for simpler skipping along compiled
4806 code. This is again a macro in internal.h so that it is next to the
4807 definition of the opcodes.
4809 34. Added support for recursive calls to individual subpatterns, along the
4810 lines of Robin Houston's patch (but implemented somewhat differently).
4812 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4813 allow it to read and process whole directories in Win32. This code was
4814 contributed by Lionel Fourquaux; it has not been tested by me.
4816 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4817 used to name a group. Names consist of alphanumerics and underscores, and must
4818 be unique. Back references use the syntax (?P=name) and recursive calls use
4819 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4820 numbers. The function pcre_fullinfo() can be used after compilation to extract
4821 a name/number map. There are three relevant calls:
4823 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4824 PCRE_INFO_NAMECOUNT yields the number of entries
4825 PCRE_INFO_NAMETABLE yields a pointer to the map.
4827 The map is a vector of fixed-size entries. The size of each entry depends on
4828 the length of the longest name used. The first two bytes of each entry are the
4829 group number, most significant byte first. There follows the corresponding
4830 name, zero terminated. The names are in alphabetical order.
4832 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4833 case instead of 255. Making it the same both with and without UTF-8 support
4834 means that the same test output works with both.
4836 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4837 calling malloc() with a zero argument.
4839 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4840 optimization. I've improved things by keeping a bitmap of backreferences with
4841 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4842 fact referenced, the optimization can be applied. It is unlikely that a
4843 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4844 the match to follow \n) will appear inside brackets with a number greater than
4845 31, but if it does, any back reference > 31 suppresses the optimization.
4847 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4848 of disabling numbered capturing parentheses. Any opening parenthesis that is
4849 not followed by ? behaves as if it were followed by ?: but named parentheses
4850 can still be used for capturing (and they will acquire numbers in the usual
4851 way).
4853 41. Redesigned the return codes from the match() function into yes/no/error so
4854 that errors can be passed back from deep inside the nested calls. A malloc
4855 failure while inside a recursive subpattern call now causes the
4856 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4858 42. It is now possible to set a limit on the number of times the match()
4859 function is called in a call to pcre_exec(). This facility makes it possible to
4860 limit the amount of recursion and backtracking, though not in a directly
4861 obvious way, because the match() function is used in a number of different
4862 circumstances. The count starts from zero for each position in the subject
4863 string (for non-anchored patterns). The default limit is, for compatibility, a
4864 large number, namely 10 000 000. You can change this in two ways:
4866 (a) When configuring PCRE before making, you can use --with-match-limit=n
4867 to set a default value for the compiled library.
4869 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4870 a different value is set. See 45 below.
4872 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4874 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4875 of things that can be changed at compile time. The first argument specifies
4876 what is wanted and the second points to where the information is to be placed.
4877 The current list of available information is:
4881 The output is an integer that is set to one if UTF-8 support is available;
4882 otherwise it is set to zero.
4886 The output is an integer that it set to the value of the code that is used for
4887 newline. It is either LF (10) or CR (13).
4891 The output is an integer that contains the number of bytes used for internal
4892 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4896 The output is an integer that contains the threshold above which the POSIX
4897 interface uses malloc() for output vectors. See item 31 above.
4901 The output is an unsigned integer that contains the default limit of the number
4902 of match() calls in a pcre_exec() execution. See 42 above.
4904 44. pcretest has been upgraded by the addition of the -C option. This causes it
4905 to extract all the available output from the new pcre_config() function, and to
4906 output it. The program then exits immediately.
4908 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4909 order to support additional features. One way would have been to define
4910 pcre_exec2() (for example) with extra arguments, but this would not have been
4911 extensible, and would also have required all calls to the original function to
4912 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4913 is used for passing in "extra" data from pcre_study().
4915 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4916 contains the following fields:
4918 flags a bitmap indicating which of the following fields are set
4919 study_data opaque data from pcre_study()
4920 match_limit a way of specifying a limit on match() calls for a specific
4921 call to pcre_exec()
4922 callout_data data for callouts (see 49 below)
4924 The flag bits are also defined in pcre.h, and are
4930 The pcre_study() function now returns one of these new pcre_extra blocks, with
4931 the actual study data pointed to by the study_data field, and the
4932 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4933 before. That is, this change is entirely upwards-compatible and requires no
4934 change to existing code.
4936 If you want to pass in additional data to pcre_exec(), you can either place it
4937 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4938 block.
4940 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4941 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4942 times with different match limits, until it finds the minimum value needed for
4943 pcre_exec() to complete. The value is then output. This can be instructive; for
4944 most simple matches the number is quite small, but for pathological cases it
4945 gets very large very quickly.
4947 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4948 returns the size of the data block pointed to by the study_data field in a
4949 pcre_extra block, that is, the value that was passed as the argument to
4950 pcre_malloc() when PCRE was getting memory in which to place the information
4951 created by pcre_study(). The fourth argument should point to a size_t variable.
4952 pcretest has been extended so that this information is shown after a successful
4953 pcre_study() call when information about the compiled regex is being displayed.
4955 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4956 because what follows is always an absolute path. (Later: it turns out that this
4957 is more than cosmetic for MinGW, because it doesn't like empty path
4958 components.)
4960 49. Some changes have been made to the callout feature (see 28 above):
4962 (i) A callout function now has three choices for what it returns:
4964 0 => success, carry on matching
4965 > 0 => failure at this point, but backtrack if possible
4966 < 0 => serious error, return this value from pcre_exec()
4968 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4969 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4970 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4971 use by callout functions. It will never be used by PCRE itself.
4973 (ii) The pcre_extra structure (see 45 above) has a void * field called
4974 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4975 pcre_callout_block structure has a field of the same name. The contents of
4976 the field passed in the pcre_extra structure are passed to the callout
4977 function in the corresponding field in the callout block. This makes it
4978 easier to use the same callout-containing regex from multiple threads. For
4979 testing, the pcretest program has a new data escape
4981 \C*n pass the number n (may be negative) as callout_data
4983 If the callout function in pcretest receives a non-zero value as
4984 callout_data, it returns that value.
4986 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4987 there were some redundant $(CFLAGS) in commands that are now specified as
4988 $(LINK), which already includes $(CFLAGS).
4990 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4991 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4992 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4993 one-byte characters throughout. Note that case-insensitive matching applies
4994 only to characters whose values are less than 256. PCRE doesn't support the
4995 notion of cases for higher-valued characters.
4997 (i) A character class whose characters are all within 0-255 is handled as
4998 a bit map, and the map is inverted for negative classes. Previously, a
4999 character > 255 always failed to match such a class; however it should
5000 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5002 (ii) A negated character class with a single character < 255 is coded as
5003 "not this character" (OP_NOT). This wasn't working properly when the test
5004 character was multibyte, either singly or repeated.
5006 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5007 mode, for example: \x{100}{2,3}.
5009 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5010 singly or repeated) now correctly test multibyte characters. However,
5011 PCRE doesn't recognize any characters with values greater than 255 as
5012 digits, spaces, or word characters. Such characters always match \D, \S,
5013 and \W, and never match \d, \s, or \w.
5015 (v) Classes may now contain characters and character ranges with values
5016 greater than 255. For example: [ab\x{100}-\x{400}].
5018 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5019 PCRE in UTF-8 mode.
5021 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5022 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5023 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5024 value.)
5026 53. The single man page has become too large. I have therefore split it up into
5027 a number of separate man pages. These also give rise to individual HTML pages;
5028 these are now put in a separate directory, and there is an index.html page that
5029 lists them all. Some hyperlinking between the pages has been installed.
5031 54. Added convenience functions for handling named capturing parentheses.
5033 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5034 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5035 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5036 are faulted.
5038 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5039 calling configure. These values are used when compiling the dftables.c program
5040 which is run to generate the source of the default character tables. They
5041 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5042 you will need to set these values.
5044 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5047 Version 3.9 02-Jan-02
5048 ---------------------
5050 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5052 2. If --disable-static was given, the building process failed when trying to
5053 build pcretest and pcregrep. (For some reason it was using libtool to compile
5054 them, which is not right, as they aren't part of the library.)
5057 Version 3.8 18-Dec-01
5058 ---------------------
5060 1. The experimental UTF-8 code was completely screwed up. It was packing the
5061 bytes in the wrong order. How dumb can you get?
5064 Version 3.7 29-Oct-01
5065 ---------------------
5067 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5068 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5069 this didn't happen under Solaris 8, where I normally test things.
5071 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5072 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5073 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5074 this fix an hour or so after the initial 3.7 release.)
5077 Version 3.6 23-Oct-01
5078 ---------------------
5080 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5081 offsets passed as NULL with zero offset count.
5083 2. The config.guess and config.sub files had not been updated when I moved to
5084 the latest autoconf.
5087 Version 3.5 15-Aug-01
5088 ---------------------
5090 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5091 had been forgotten.
5093 2. By using declared but undefined structures, we can avoid using "void"
5094 definitions in pcre.h while keeping the internal definitions of the structures
5095 private.
5097 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5098 user point of view, this means that both static and shared libraries are built
5099 by default, but this can be individually controlled. More of the work of
5100 handling this static/shared cases is now inside libtool instead of PCRE's make
5101 file.
5103 4. The pcretest utility is now installed along with pcregrep because it is
5104 useful for users (to test regexs) and by doing this, it automatically gets
5105 relinked by libtool. The documentation has been turned into a man page, so
5106 there are now .1, .txt, and .html versions in /doc.
5108 5. Upgrades to pcregrep:
5109 (i) Added long-form option names like gnu grep.
5110 (ii) Added --help to list all options with an explanatory phrase.
5111 (iii) Added -r, --recursive to recurse into sub-directories.
5112 (iv) Added -f, --file to read patterns from a file.
5114 6. pcre_exec() was referring to its "code" argument before testing that
5115 argument for NULL (and giving an error if it was NULL).
5117 7. Upgraded Makefile.in to allow for compiling in a different directory from
5118 the source directory.
5120 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5121 options bits, the pointer it was passed was to an int instead of to an unsigned
5122 long int. This mattered only on 64-bit systems.
5124 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5125 generated) instead of pcre.in, which it its source. Also made the same change
5126 in several of the .c files.
5128 10. A new release of gcc defines printf() as a macro, which broke pcretest
5129 because it had an ifdef in the middle of a string argument for printf(). Fixed
5130 by using separate calls to printf().
5132 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5133 script, to force use of CR or LF instead of \n in the source. On non-Unix
5134 systems, the value can be set in config.h.
5136 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5137 absolute limit. Changed the text of the error message to make this clear, and
5138 likewise updated the man page.
5140 13. The limit of 99 on the number of capturing subpatterns has been removed.
5141 The new limit is 65535, which I hope will not be a "real" limit.
5144 Version 3.4 22-Aug-00
5145 ---------------------
5147 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5149 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5152 Version 3.3 01-Aug-00
5153 ---------------------
5155 1. If an octal character was given, but the value was greater than \377, it
5156 was not getting masked to the least significant bits, as documented. This could
5157 lead to crashes in some systems.
5159 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5160 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5162 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5163 These just pass their arguments on to (pcre_free)(), but they are provided
5164 because some uses of PCRE bind it to non-C systems that can call its functions,
5165 but cannot call free() or pcre_free() directly.
5167 4. Add "make test" as a synonym for "make check". Corrected some comments in
5168 the Makefile.
5170 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5171 Makefile.
5173 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5174 command called pgrep for grepping around the active processes.
5176 7. Added the beginnings of support for UTF-8 character strings.
5178 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5179 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5180 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5181 out for the ar command.)
5184 Version 3.2 12-May-00
5185 ---------------------
5187 This is purely a bug fixing release.
5189 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5190 of ZA. This was just one example of several cases that could provoke this bug,
5191 which was introduced by change 9 of version 2.00. The code for breaking
5192 infinite loops after an iteration that matches an empty string was't working
5193 correctly.
5195 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5196 when matched against abbab (for example). After matching an empty string, it
5197 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5198 caused it to match further down the string than it should.
5200 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5201 was there because it doesn't seem to be needed, and it causes trouble on some
5202 systems, as it is not a Standard C header. It has been removed.
5204 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5205 were reported on the Macintosh. The changes were from
5207 while ((c = *(++ptr)) != 0 && c != '\n');
5208 to
5209 while ((c = *(++ptr)) != 0 && c != '\n') ;
5211 Totally extraordinary, but if that's what it takes...
5213 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5214 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5215 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5216 assumes the way PCRE uses memmove() (always moving upwards).
5218 6. PCRE is being used in one environment where strchr() is not available. There
5219 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5220 faster code anyway.
5223 Version 3.1 09-Feb-00
5224 ---------------------
5226 The only change in this release is the fixing of some bugs in Makefile.in for
5227 the "install" target:
5229 (1) It was failing to install pcreposix.h.
5231 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5234 Version 3.0 01-Feb-00
5235 ---------------------
5237 1. Add support for the /+ modifier to perltest (to output $` like it does in
5238 pcretest).
5240 2. Add support for the /g modifier to perltest.
5242 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5243 matches null strings.
5245 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5246 pattern. Perl treats empty patterns specially - it reuses the most recent
5247 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5248 effect.
5250 5. The POSIX interface was broken in that it was just handing over the POSIX
5251 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5252 required a bigger vector, with some working space on the end. This means that
5253 the POSIX wrapper now has to get and free some memory, and copy the results.
5255 6. Added some simple autoconf support, placing the test data and the
5256 documentation in separate directories, re-organizing some of the
5257 information files, and making it build pcre-config (a GNU standard). Also added
5258 libtool support for building PCRE as a shared library, which is now the
5259 default.
5261 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5262 09 are not valid octal constants. Single digits will be used for minor values
5263 less than 10.
5265 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5266 existing programs that set these in the POSIX interface can use PCRE without
5267 modification.
5269 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5270 return all that pcre_info() returns, plus additional data. The pcre_info()
5271 function is retained for compatibility, but is considered to be obsolete.
5273 10. Added experimental recursion feature (?R) to handle one common case that
5274 Perl 5.6 will be able to do with (?p{...}).
5276 11. Added support for POSIX character classes like [:alpha:], which Perl is
5277 adopting.
5280 Version 2.08 31-Aug-99
5281 ----------------------
5283 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5284 trying to match at the startoffset position, but instead was moving forward to
5285 the next newline as if a previous match had failed.
5287 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5288 and could get into a loop if a null string was matched other than at the start
5289 of the subject.
5291 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5292 be distinguished at compile time, and for completeness also added PCRE_DATE.
5294 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5295 in GnuWin32 environments.
5298 Version 2.07 29-Jul-99
5299 ----------------------
5301 1. The documentation is now supplied in plain text form and HTML as well as in
5302 the form of man page sources.
5304 2. C++ compilers don't like assigning (void *) values to other pointer types.
5305 In particular this affects malloc(). Although there is no problem in Standard
5306 C, I've put in casts to keep C++ compilers happy.
5308 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5309 should be (const char *).
5311 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5312 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5313 However, I haven't made this a standard facility. The documentation doesn't
5314 mention it, and the Makefile doesn't support it.
5316 5. The Makefile now contains an "install" target, with editable destinations at
5317 the top of the file. The pcretest program is not installed.
5319 6. pgrep -V now gives the PCRE version number and date.
5321 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5322 causing the entire string to be ignored, instead of just the last character.
5324 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5325 non-matching string, it can take a very, very long time, even for strings of
5326 quite modest length, because of the nested recursion. PCRE now does better in
5327 some of these cases. It does this by remembering the last required literal
5328 character in the pattern, and pre-searching the subject to ensure it is present
5329 before running the real match. In other words, it applies a heuristic to detect
5330 some types of certain failure quickly, and in the above example, if presented
5331 with a string that has no trailing " it gives "no match" very quickly.
5333 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5334 other alternatives are tried instead.
5337 Version 2.06 09-Jun-99
5338 ----------------------
5340 1. Change pcretest's output for amount of store used to show just the code
5341 space, because the remainder (the data block) varies in size between 32-bit and
5342 64-bit systems.
5344 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5345 start matching at. This allows lookbehinds to work when searching for multiple
5346 occurrences in a string.
5348 3. Added additional options to pcretest for testing multiple occurrences:
5350 /+ outputs the rest of the string that follows a match
5351 /g loops for multiple occurrences, using the new startoffset argument
5352 /G loops for multiple occurrences by passing an incremented pointer
5354 4. PCRE wasn't doing the "first character" optimization for patterns starting
5355 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5356 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5357 the letter 'x'. On long subject strings, this gives a significant speed-up.
5360 Version 2.05 21-Apr-99
5361 ----------------------
5363 1. Changed the type of magic_number from int to long int so that it works
5364 properly on 16-bit systems.
5366 2. Fixed a bug which caused patterns starting with .* not to work correctly
5367 when the subject string contained newline characters. PCRE was assuming
5368 anchoring for such patterns in all cases, which is not correct because .* will
5369 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5370 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5371 must be retried after every newline in the subject.
5374 Version 2.04 18-Feb-99
5375 ----------------------
5377 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5378 computation of the store needed to hold the pattern was incorrect (too large).
5379 If such patterns were nested a few deep, this could multiply and become a real
5380 problem.
5382 2. Added /M option to pcretest to show the memory requirement of a specific
5383 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5385 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5386 compiled in such a way that the backtracking after subsequent failure was
5387 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5388 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5391 Version 2.03 02-Feb-99
5392 ----------------------
5394 1. Fixed typo and small mistake in man page.
5396 2. Added 4th condition (GPL supersedes if conflict) and created separate
5397 LICENCE file containing the conditions.
5399 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5400 Perl, that is the internal \ allows the delimiter to be included in the
5401 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5402 the final delimiter, add \ to the end of the pattern (to test the error).
5404 4. Added the convenience functions for extracting substrings after a successful
5405 match. Updated pcretest to make it able to test these functions.
5408 Version 2.02 14-Jan-99
5409 ----------------------
5411 1. Initialized the working variables associated with each extraction so that
5412 their saving and restoring doesn't refer to uninitialized store.
5414 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5415 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5416 fix the problem.
5418 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5419 calls, and wasn't printing the correct value for compiling calls. Increased the
5420 default value of LOOPREPEAT, and the number of significant figures in the
5421 times.
5423 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5425 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5426 a building problem on Windows NT with a FAT file system.
5429 Version 2.01 21-Oct-98
5430 ----------------------
5432 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5433 to character tables built by pcre_maketables() in the current locale. If NULL
5434 is passed, the default tables are used.
5437 Version 2.00 24-Sep-98
5438 ----------------------
5440 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5441 it any more.
5443 2. Allow quantification of (?>) groups, and make it work correctly.
5445 3. The first character computation wasn't working for (?>) groups.
5447 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5448 end of the subject) and add 5.005's \z, which really does match only at the
5449 very end of the subject.
5451 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5453 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5454 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5455 localized options. All options to pcre_study() were also removed.
5457 7. Add other new features from 5.005:
5459 $(?<= positive lookbehind
5460 $(?<! negative lookbehind
5461 (?imsx-imsx) added the unsetting capability
5462 such a setting is global if at outer level; local otherwise
5463 (?imsx-imsx:) non-capturing groups with option setting
5464 (?(cond)re|re) conditional pattern matching
5466 A backreference to itself in a repeated group matches the previous
5467 captured string.
5469 8. General tidying up of studying (both automatic and via "study")
5470 consequential on the addition of new assertions.
5472 9. As in 5.005, unlimited repeated groups that could match an empty substring
5473 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5474 runtime if any iteration does actually match an empty substring.
5476 10. Include the RunTest script in the distribution.
5478 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5479 discrepancies, some of which were old and were also with respect to 5.004. They
5480 have now been fixed.
5483 Version 1.09 28-Apr-98
5484 ----------------------
5486 1. A negated single character class followed by a quantifier with a minimum
5487 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5488 program crashes, or just wrong answers. This did not apply to negated classes
5489 containing more than one character, or to minima other than one.
5492 Version 1.08 27-Mar-98
5493 ----------------------
5495 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5497 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5498 latter must appear before anything that relies on it in the pattern.
5501 Version 1.07 16-Feb-98
5502 ----------------------
5504 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5505 repeat of a potentially empty string).
5508 Version 1.06 23-Jan-98
5509 ----------------------
5511 1. Added Markus Oberhumer's little patches for C++.
5513 2. Literal strings longer than 255 characters were broken.
5516 Version 1.05 23-Dec-97
5517 ----------------------
5519 1. Negated character classes containing more than one character were failing if
5520 PCRE_CASELESS was set at run time.
5523 Version 1.04 19-Dec-97
5524 ----------------------
5526 1. Corrected the man page, where some "const" qualifiers had been omitted.
5528 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5529 input syntax.
5531 3. Fixed memory leak which occurred when a regex with back references was
5532 matched with an offsets vector that wasn't big enough. The temporary memory
5533 that is used in this case wasn't being freed if the match failed.
5535 4. Tidied pcretest to ensure it frees memory that it gets.
5537 5. Temporary memory was being obtained in the case where the passed offsets
5538 vector was exactly big enough.
5540 6. Corrected definition of offsetof() from change 5 below.
5542 7. I had screwed up change 6 below and broken the rules for the use of
5543 setjmp(). Now fixed.
5546 Version 1.03 18-Dec-97
5547 ----------------------
5549 1. A erroneous regex with a missing opening parenthesis was correctly
5550 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5551 on some systems.
5553 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5554 it was reported that one broken compiler failed on the former because "code" is
5555 also an independent variable.
5557 3. The erroneous regex a[]b caused an array overrun reference.
5559 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5560 fail on data ending with that character. (It was going on too far, and checking
5561 the next character, typically a binary zero.) This was specific to the
5562 optimized code for single-character negative classes.
5564 5. Added a contributed patch from the TIN world which does the following:
5566 + Add an undef for memmove, in case the the system defines a macro for it.
5568 + Add a definition of offsetof(), in case there isn't one. (I don't know
5569 the reason behind this - offsetof() is part of the ANSI standard - but
5570 it does no harm).
5572 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5573 most of the places where whitespace preceded '#'. I have given up and
5574 allowed the remaining 2 cases to be at the margin.
5576 + Rename some variables in pcre to eliminate shadowing. This seems very
5577 pedantic, but does no harm, of course.
5579 6. Moved the call to setjmp() into its own function, to get rid of warnings
5580 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5582 7. Constructs such as \d{8,} were compiling into the equivalent of
5583 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5584 outcome, but in this particular case used more store than had been allocated,
5585 which caused the bug to be discovered because it threw up an internal error.
5587 8. The debugging code in both pcre and pcretest for outputting the compiled
5588 form of a regex was going wrong in the case of back references followed by
5589 curly-bracketed repeats.
5592 Version 1.02 12-Dec-97
5593 ----------------------
5595 1. Typos in pcre.3 and comments in the source fixed.
5597 2. Applied a contributed patch to get rid of places where it used to remove
5598 'const' from variables, and fixed some signed/unsigned and uninitialized
5599 variable warnings.
5601 3. Added the "runtest" target to Makefile.
5603 4. Set default compiler flag to -O2 rather than just -O.
5606 Version 1.01 19-Nov-97
5607 ----------------------
5609 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5610 like /([ab]*)*/, that is, for classes with more than one character in them.
5612 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5613 as /((?>a*))*/ (a PCRE_EXTRA facility).
5616 Version 1.00 18-Nov-97
5617 ----------------------
5619 1. Added compile-time macros to support systems such as SunOS4 which don't have
5620 memmove() or strerror() but have other things that can be used instead.
5622 2. Arranged that "make clean" removes the executables.
5625 Version 0.99 27-Oct-97
5626 ----------------------
5628 1. Fixed bug in code for optimizing classes with only one character. It was
5629 initializing a 32-byte map regardless, which could cause it to run off the end
5630 of the memory it had got.
5632 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5635 Version 0.98 22-Oct-97
5636 ----------------------
5638 1. Fixed bug in code for handling temporary memory usage when there are more
5639 back references than supplied space in the ovector. This could cause segfaults.
5642 Version 0.97 21-Oct-97
5643 ----------------------
5645 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5647 2. Optimized negated single characters not to use a bit map.
5649 3. Brought error texts together as macro definitions; clarified some of them;
5650 fixed one that was wrong - it said "range out of order" when it meant "invalid
5651 escape sequence".
5653 4. Changed some char * arguments to const char *.
5655 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5657 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5658 pcretest.
5661 Version 0.96 16-Oct-97
5662 ----------------------
5664 1. Added a simple "pgrep" utility to the distribution.
5666 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5667 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5668 where "ddd" means "one or more decimal digits".
5670 3. Fixed serious bug. If a pattern had a back reference, but the call to
5671 pcre_exec() didn't supply a large enough ovector to record the related
5672 identifying subpattern, the match always failed. PCRE now remembers the number
5673 of the largest back reference, and gets some temporary memory in which to save
5674 the offsets during matching if necessary, in order to ensure that
5675 backreferences always work.
5677 4. Increased the compatibility with Perl in a number of ways:
5679 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5680 to request this handling. The option can be set at compile or exec time.
5682 (b) $ matches before a terminating newline by default; an option
5683 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5684 mode). The option can be set at compile or exec time.
5686 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5687 the same as Perl's. If the decimal number it represents is less than 10
5688 or there aren't that many previous left capturing parentheses, an octal
5689 escape is read. Inside a character class, it's always an octal escape,
5690 even if it is a single digit.
5692 (d) An escaped but undefined alphabetic character is taken as a literal,
5693 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5694 escapes.
5696 (e) {0} is now permitted. (The previous item is removed from the compiled
5697 pattern).
5699 5. Changed all the names of code files so that the basic parts are no longer
5700 than 10 characters, and abolished the teeny "globals.c" file.
5702 6. Changed the handling of character classes; they are now done with a 32-byte
5703 bit map always.
5705 7. Added the -d and /D options to pcretest to make it possible to look at the
5706 internals of compilation without having to recompile pcre.
5709 Version 0.95 23-Sep-97
5710 ----------------------
5712 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5713 \x20 at the start of a run of normal characters. These were being treated as
5714 real characters, instead of the source characters being re-checked.
5717 Version 0.94 18-Sep-97
5718 ----------------------
5720 1. The functions are now thread-safe, with the caveat that the global variables
5721 containing pointers to malloc() and free() or alternative functions are the
5722 same for all threads.
5724 2. Get pcre_study() to generate a bitmap of initial characters for non-
5725 anchored patterns when this is possible, and use it if passed to pcre_exec().
5728 Version 0.93 15-Sep-97
5729 ----------------------
5731 1. /(b)|(:+)/ was computing an incorrect first character.
5733 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5734 but not actually doing anything yet.
5736 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5737 as Perl does (e.g. [-az] or [az-]).
5739 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5740 all possible positions.
5742 5. Split up into different modules to avoid including unneeded functions in a
5743 compiled binary. However, compile and exec are still in one module. The "study"
5744 function is split off.
5746 6. The character tables are now in a separate module whose source is generated
5747 by an auxiliary program - but can then be edited by hand if required. There are
5748 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5749 toupper() in the code.
5751 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5752 make them global. Abolish the function for setting them, as the caller can now
5753 set them directly.
5756 Version 0.92 11-Sep-97
5757 ----------------------
5759 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5760 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5762 2. Caseless matching was not working in character classes if the characters in
5763 the pattern were in upper case.
5765 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5767 4. Make PCRE_ANCHORED public and accept as a compile option.
5769 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5770 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5771 pass them.
5773 6. Give an error if bad option bits passed at compile or run time.
5775 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5776 pcretest to cause it to pass that flag.
5778 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5779 options, and the first character, if set.
5781 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5784 Version 0.91 10-Sep-97
5785 ----------------------
5787 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5788 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5790 2. PCRE was looping on encountering an indefinitely repeated back reference to
5791 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5792 Perl does - treats the match as successful.
5794 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5