ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1569 - (show annotations)
Fri Jun 19 16:10:07 2015 UTC (5 years, 10 months ago) by ph10
File size: 280607 byte(s)
Recognize EBCDIC non-breaking space and give error for \p in a class when no 
UCP support.
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|)+)+$()/.
55 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
56 \a and \e in test subject lines.
58 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
59 instead of the EBCDIC value.
61 14. The handling of \c in an EBCDIC environment has been revised so that it is
62 now compatible with the specification in Perl's perlebcdic page.
64 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
65 ASCII/Unicode. This has now been added to the list of characters that are
66 recognized as white space in EBCDIC.
68 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
69 error (correctly) when used outside a class, but did not give an error
70 within a class.
73 Version 8.37 28-April-2015
74 --------------------------
76 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
77 for those parentheses to be closed with whatever has been captured so far.
78 However, it was failing to mark any other groups between the hightest
79 capture so far and the currrent group as "unset". Thus, the ovector for
80 those groups contained whatever was previously there. An example is the
81 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
83 2. If an assertion condition was quantified with a minimum of zero (an odd
84 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
86 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
87 unrecognized modifier, a crash could occur.
89 4. An attempt to do global matching in pcretest with a zero-length ovector
90 caused a crash.
92 5. Fixed a memory leak during matching that could occur for a subpattern
93 subroutine call (recursive or otherwise) if the number of captured groups
94 that had to be saved was greater than ten.
96 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
97 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
98 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
100 7. A UTF pattern containing a "not" match of a non-ASCII character and a
101 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
103 8. When a pattern is compiled, it remembers the highest back reference so that
104 when matching, if the ovector is too small, extra memory can be obtained to
105 use instead. A conditional subpattern whose condition is a check on a
106 capture having happened, such as, for example in the pattern
107 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
108 setting the highest backreference number. This mattered only if pcre_exec()
109 was called with an ovector that was too small to hold the capture, and there
110 was no other kind of back reference (a situation which is probably quite
111 rare). The effect of the bug was that the condition was always treated as
112 FALSE when the capture could not be consulted, leading to a incorrect
113 behaviour by pcre_exec(). This bug has been fixed.
115 9. A reference to a duplicated named group (either a back reference or a test
116 for being set in a conditional) that occurred in a part of the pattern where
117 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
118 to be incorrectly calculated, leading to overwriting.
120 10. A mutually recursive set of back references such as (\2)(\1) caused a
121 segfault at study time (while trying to find the minimum matching length).
122 The infinite loop is now broken (with the minimum length unset, that is,
123 zero).
125 11. If an assertion that was used as a condition was quantified with a minimum
126 of zero, matching went wrong. In particular, if the whole group had
127 unlimited repetition and could match an empty string, a segfault was
128 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
129 assertions to be quantified, but not if they are being used as conditions,
130 so the above pattern is faulted by Perl. PCRE has now been changed so that
131 it also rejects such patterns.
133 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
134 failed to allow the zero-repeat case if pcre2_exec() was called with an
135 ovector too small to capture the group.
137 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
138 Red Hat Product Security:
140 (a) A crash if /K and /F were both set with the option to save the compiled
141 pattern.
143 (b) Another crash if the option to print captured substrings in a callout
144 was combined with setting a null ovector, for example \O\C+ as a subject
145 string.
147 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
148 forward reference repeated a large (but limited) number of times within a
149 repeated outer group that has a zero minimum quantifier, caused incorrect
150 code to be compiled, leading to the error "internal error:
151 previously-checked referenced subpattern not found" when an incorrect
152 memory address was read. This bug was reported as "heap overflow",
153 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
154 CVE-2015-2325.
156 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
157 call within a group that also contained a recursive back reference caused
158 incorrect code to be compiled. This bug was reported as "heap overflow",
159 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
160 number CVE-2015-2326.
162 24. Computing the size of the JIT read-only data in advance has been a source
163 of various issues, and new ones are still appear unfortunately. To fix
164 existing and future issues, size computation is eliminated from the code,
165 and replaced by on-demand memory allocation.
167 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
168 adjacent to the end of the range, and the range contained characters with
169 more than one other case, caused incorrect behaviour when compiled in UTF
170 mode. In that example, the range a-j was left out of the class.
172 26. Fix JIT compilation of conditional blocks, which assertion
173 is converted to (*FAIL). E.g: /(?(?!))/.
175 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
176 discovered by the LLVM fuzzer.
178 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
179 when this assertion was used as a condition, for example (?(?!)a|b). In
180 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
181 error about an unsupported item.
183 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
184 possessification code could take exponential time to complete. A recursion
185 depth limit of 1000 has been imposed to limit the resources used by this
186 optimization.
188 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
189 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
190 because \S ensures they are all in the class. The code for doing this was
191 interacting badly with the code for computing the amount of space needed to
192 compile the pattern, leading to a buffer overflow. This bug was discovered
193 by the LLVM fuzzer.
195 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
196 other kinds of group caused stack overflow at compile time. This bug was
197 discovered by the LLVM fuzzer.
199 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
200 between a subroutine call and its quantifier was incorrectly compiled,
201 leading to buffer overflow or other errors. This bug was discovered by the
202 LLVM fuzzer.
204 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
205 assertion after (?(. The code was failing to check the character after
206 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
207 was discovered by the LLVM fuzzer.
209 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
210 a fixed maximum following a group that contains a subroutine reference was
211 incorrectly compiled and could trigger buffer overflow. This bug was
212 discovered by the LLVM fuzzer.
214 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
215 caused a stack overflow instead of the diagnosis of a non-fixed length
216 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
218 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
219 (e.g. /(?<=\Ka)/) could make pcregrep loop.
221 37. There was a similar problem to 36 in pcretest for global matches.
223 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
224 and a subsequent item in the pattern caused a non-match, backtracking over
225 the repeated \X did not stop, but carried on past the start of the subject,
226 causing reference to random memory and/or a segfault. There were also some
227 other cases where backtracking after \C could crash. This set of bugs was
228 discovered by the LLVM fuzzer.
230 39. The function for finding the minimum length of a matching string could take
231 a very long time if mutual recursion was present many times in a pattern,
232 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
233 method has been implemented. This infelicity was discovered by the LLVM
234 fuzzer.
236 40. Static linking against the PCRE library using the pkg-config module was
237 failing on missing pthread symbols.
240 Version 8.36 26-September-2014
241 ------------------------------
243 1. Got rid of some compiler warnings in the C++ modules that were shown up by
244 -Wmissing-field-initializers and -Wunused-parameter.
246 2. The tests for quantifiers being too big (greater than 65535) were being
247 applied after reading the number, and stupidly assuming that integer
248 overflow would give a negative number. The tests are now applied as the
249 numbers are read.
251 3. Tidy code in pcre_exec.c where two branches that used to be different are
252 now the same.
254 4. The JIT compiler did not generate match limit checks for certain
255 bracketed expressions with quantifiers. This may lead to exponential
256 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
257 issue should be resolved now.
259 5. Fixed an issue, which occures when nested alternatives are optimized
260 with table jumps.
262 6. Inserted two casts and changed some ints to size_t in the light of some
263 reported 64-bit compiler warnings (Bugzilla 1477).
265 7. Fixed a bug concerned with zero-minimum possessive groups that could match
266 an empty string, which sometimes were behaving incorrectly in the
267 interpreter (though correctly in the JIT matcher). This pcretest input is
268 an example:
270 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
273 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
274 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
275 for an empty string was breaking the inner loop and carrying on at a lower
276 level, when possessive repeated groups should always return to a higher
277 level as they have no backtrack points in them. The empty string test now
278 occurs at the outer level.
280 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
281 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
283 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
284 doesn't).
286 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
287 to hold a pattern to be incorrectly computed (too small) when there were
288 named back references to duplicated names. This could cause "internal
289 error: code overflow" or "double free or corruption" or other memory
290 handling errors.
292 11. When named subpatterns had the same prefixes, back references could be
293 confused. For example, in this pattern:
295 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
297 the reference to 'Name' was incorrectly treated as a reference to a
298 duplicate name.
300 12. A pattern such as /^s?c/mi8 where the optional character has more than
301 one "other case" was incorrectly compiled such that it would only try to
302 match starting at "c".
304 13. When a pattern starting with \s was studied, VT was not included in the
305 list of possible starting characters; this should have been part of the
306 8.34/18 patch.
308 14. If a character class started [\Qx]... where x is any character, the class
309 was incorrectly terminated at the ].
311 15. If a pattern that started with a caseless match for a character with more
312 than one "other case" was studied, PCRE did not set up the starting code
313 unit bit map for the list of possible characters. Now it does. This is an
314 optimization improvement, not a bug fix.
316 16. The Unicode data tables have been updated to Unicode 7.0.0.
318 17. Fixed a number of memory leaks in pcregrep.
320 18. Avoid a compiler warning (from some compilers) for a function call with
321 a cast that removes "const" from an lvalue by using an intermediate
322 variable (to which the compiler does not object).
324 19. Incorrect code was compiled if a group that contained an internal recursive
325 back reference was optional (had quantifier with a minimum of zero). This
326 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
327 caused segmentation faults because of stack overflows at compile time.
329 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
330 group that is quantified with an indefinite repeat, caused a compile-time
331 loop which used up all the system stack and provoked a segmentation fault.
332 This was not the same bug as 19 above.
334 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
335 Patch by Mike Frysinger.
338 Version 8.35 04-April-2014
339 --------------------------
341 1. A new flag is set, when property checks are present in an XCLASS.
342 When this flag is not set, PCRE can perform certain optimizations
343 such as studying these XCLASS-es.
345 2. The auto-possessification of character sets were improved: a normal
346 and an extended character set can be compared now. Furthermore
347 the JIT compiler optimizes more character set checks.
349 3. Got rid of some compiler warnings for potentially uninitialized variables
350 that show up only when compiled with -O2.
352 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
353 of a match later then the end of the match. The pcretest program was not
354 handling the case sensibly - it was outputting from the start to the next
355 binary zero. It now reports this situation in a message, and outputs the
356 text from the end to the start.
358 5. Fast forward search is improved in JIT. Instead of the first three
359 characters, any three characters with fixed position can be searched.
360 Search order: first, last, middle.
362 6. Improve character range checks in JIT. Characters are read by an inprecise
363 function now, which returns with an unknown value if the character code is
364 above a certain threshold (e.g: 256). The only limitation is that the value
365 must be bigger than the threshold as well. This function is useful when
366 the characters above the threshold are handled in the same way.
368 7. The macros whose names start with RAWUCHAR are placeholders for a future
369 mode in which only the bottom 21 bits of 32-bit data items are used. To
370 make this more memorable for those maintaining the code, the names have
371 been changed to start with UCHAR21, and an extensive comment has been added
372 to their definition.
374 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
375 to the export list in Makefile.am (they were accidentally omitted from the
376 8.34 tarball).
378 9. The informational output from pcretest used the phrase "starting byte set"
379 which is inappropriate for the 16-bit and 32-bit libraries. As the output
380 for "first char" and "need char" really means "non-UTF-char", I've changed
381 "byte" to "char", and slightly reworded the output. The documentation about
382 these values has also been (I hope) clarified.
384 10. Another JIT related optimization: use table jumps for selecting the correct
385 backtracking path, when more than four alternatives are present inside a
386 bracket.
388 11. Empty match is not possible, when the minimum length is greater than zero,
389 and there is no \K in the pattern. JIT should avoid empty match checks in
390 such cases.
392 12. In a caseless character class with UCP support, when a character with more
393 than one alternative case was not the first character of a range, not all
394 the alternative cases were added to the class. For example, s and \x{17f}
395 are both alternative cases for S: the class [RST] was handled correctly,
396 but [R-T] was not.
398 13. The configure.ac file always checked for pthread support when JIT was
399 enabled. This is not used in Windows, so I have put this test inside a
400 check for the presence of windows.h (which was already tested for).
402 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
403 The algorithm provides a way to skip certain starting offsets, and usually
404 faster than linear prefix searches.
406 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
407 as for 'fr_FR' and 'french'. For some reason, however, it then used the
408 Windows-specific input and output files, which have 'french' screwed in.
409 So this could never have worked. One of the problems with locales is that
410 they aren't always the same. I have now updated RunTest so that it checks
411 the output of the locale test (test 3) against three different output
412 files, and it allows the test to pass if any one of them matches. With luck
413 this should make the test pass on some versions of Solaris where it was
414 failing. Because of the uncertainty, the script did not used to stop if
415 test 3 failed; it now does. If further versions of a French locale ever
416 come to light, they can now easily be added.
418 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
419 there was a message during ./configure, but it did not stop. This now
420 provokes an error. The invalid example in README has been corrected.
421 If a value less than the minimum is given, the minimum value has always
422 been used, but now a warning is given.
424 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
425 was a bug in the test system, which is now fixed. Also, the list of various
426 configurations that are tested for each release did not have one with both
427 16/32 bits and --enable-bar-anycrlf. It now does.
429 18. pcretest was missing "-C bsr" for displaying the \R default setting.
431 19. Little endian PowerPC systems are supported now by the JIT compiler.
433 20. The fast forward newline mechanism could enter to an infinite loop on
434 certain invalid UTF-8 input. Although we don't support these cases
435 this issue can be fixed by a performance optimization.
437 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
438 not take account if existing stack usage. There is now a new global
439 variable called pcre_stack_guard that can be set to point to an external
440 function to check stack availability. It is called at the start of
441 processing every parenthesized group.
443 22. A typo in the code meant that in ungreedy mode the max/min qualifier
444 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
445 not match "ab".
447 23. When UTF was disabled, the JIT program reported some incorrect compile
448 errors. These messages are silenced now.
450 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
451 compiler.
453 25. Change all the temporary files used in RunGrepTest to be different to those
454 used by RunTest so that the tests can be run simultaneously, for example by
455 "make -j check".
458 Version 8.34 15-December-2013
459 -----------------------------
461 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
462 executable memory. Patch inspired by Carsten Klein.
464 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
465 this macro is never tested and has no effect, because the work to support
466 coverage involves only compiling and linking options and special targets in
467 the Makefile. The comment in config.h implied that defining the macro would
468 enable coverage support, which is totally false. There was also support for
469 setting this macro in the CMake files (my fault, I just copied it from
470 configure). SUPPORT_GCOV has now been removed.
472 3. Make a small performance improvement in strlen16() and strlen32() in
473 pcretest.
475 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
476 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
477 these cases). There was also one in pcretest.c.
479 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
481 6. In UTF mode, the code for checking whether a group could match an empty
482 string (which is used for indefinitely repeated groups to allow for
483 breaking an infinite loop) was broken when the group contained a repeated
484 negated single-character class with a character that occupied more than one
485 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
486 UTF-8 mode). The effect was undefined: the group might or might not be
487 deemed as matching an empty string, or the program might have crashed.
489 7. The code for checking whether a group could match an empty string was not
490 recognizing that \h, \H, \v, \V, and \R must match a character.
492 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
493 an empty string. If it can, pcretest shows this in its information output.
495 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
496 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
497 matched by pcre_exec() without using JIT:
499 (a) If the rest of the pattern did not match after a maximal run of
500 grapheme clusters, the code for backing up to try with fewer of them
501 did not always back up over a full grapheme when characters that do not
502 have the modifier quality were involved, e.g. Hangul syllables.
504 (b) If the match point in a subject started with modifier character, and
505 there was no match, the code could incorrectly back up beyond the match
506 point, and potentially beyond the first character in the subject,
507 leading to a segfault or an incorrect match result.
509 10. A conditional group with an assertion condition could lead to PCRE
510 recording an incorrect first data item for a match if no other first data
511 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
512 first data item, and therefore matched "ca" after "c" instead of at the
513 start.
515 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
516 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
518 12. The source of pcregrep now includes z/OS-specific code so that it can be
519 compiled for z/OS as part of the special z/OS distribution.
521 13. Added the -T and -TM options to pcretest.
523 14. The code in pcre_compile.c for creating the table of named capturing groups
524 has been refactored. Instead of creating the table dynamically during the
525 actual compiling pass, the information is remembered during the pre-compile
526 pass (on the stack unless there are more than 20 named groups, in which
527 case malloc() is used) and the whole table is created before the actual
528 compile happens. This has simplified the code (it is now nearly 150 lines
529 shorter) and prepared the way for better handling of references to groups
530 with duplicate names.
532 15. A back reference to a named subpattern when there is more than one of the
533 same name now checks them in the order in which they appear in the pattern.
534 The first one that is set is used for the reference. Previously only the
535 first one was inspected. This change makes PCRE more compatible with Perl.
537 16. Unicode character properties were updated from Unicode 6.3.0.
539 17. The compile-time code for auto-possessification has been refactored, based
540 on a patch by Zoltan Herczeg. It now happens after instead of during
541 compilation. The code is cleaner, and more cases are handled. The option
542 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
543 options in pcretest are provided to set it. It can also be set by
544 (*NO_AUTO_POSSESS) at the start of a pattern.
546 18. The character VT has been added to the default ("C" locale) set of
547 characters that match \s and are generally treated as white space,
548 following this same change in Perl 5.18. There is now no difference between
549 "Perl space" and "POSIX space". Whether VT is treated as white space in
550 other locales depends on the locale.
552 19. The code for checking named groups as conditions, either for being set or
553 for being recursed, has been refactored (this is related to 14 and 15
554 above). Processing unduplicated named groups should now be as fast at
555 numerical groups, and processing duplicated groups should be faster than
556 before.
558 20. Two patches to the CMake build system, by Alexander Barkov:
560 (1) Replace the "source" command by "." in CMakeLists.txt because
561 "source" is a bash-ism.
563 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
564 without these the CMake build does not work on Solaris.
566 21. Perl has changed its handling of \8 and \9. If there is no previously
567 encountered capturing group of those numbers, they are treated as the
568 literal characters 8 and 9 instead of a binary zero followed by the
569 literals. PCRE now does the same.
571 22. Following Perl, added \o{} to specify codepoints in octal, making it
572 possible to specify values greater than 0777 and also making them
573 unambiguous.
575 23. Perl now gives an error for missing closing braces after \x{... instead of
576 treating the string as literal. PCRE now does the same.
578 24. RunTest used to grumble if an inappropriate test was selected explicitly,
579 but just skip it when running all tests. This make it awkward to run ranges
580 of tests when one of them was inappropriate. Now it just skips any
581 inappropriate tests, as it always did when running all tests.
583 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
584 character types such as \d or \w, too many callouts were inserted, and the
585 data that they returned was rubbish.
587 26. In UCP mode, \s was not matching two of the characters that Perl matches,
588 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
589 were matched by \h. The code has now been refactored so that the lists of
590 the horizontal and vertical whitespace characters used for \h and \v (which
591 are defined only in one place) are now also used for \s.
593 27. Add JIT support for the 64 bit TileGX architecture.
594 Patch by Jiong Wang (Tilera Corporation).
596 28. Possessive quantifiers for classes (both explicit and automatically
597 generated) now use special opcodes instead of wrapping in ONCE brackets.
599 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
600 (because it's meaningless), this was not happening when PCRE_CASELESS was
601 set. Not wrong, but inefficient.
603 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
604 properties for \w, \d, etc) is present in a test regex. Otherwise if the
605 test contains no characters greater than 255, Perl doesn't realise it
606 should be using Unicode semantics.
608 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
609 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
610 does in Unicode mode.
612 32. Added the "forbid" facility to pcretest so that putting tests into the
613 wrong test files can sometimes be quickly detected.
615 33. There is now a limit (default 250) on the depth of nesting of parentheses.
616 This limit is imposed to control the amount of system stack used at compile
617 time. It can be changed at build time by --with-parens-nest-limit=xxx or
618 the equivalent in CMake.
620 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
621 errors. Perl warns for these when in warning mode, but PCRE has no facility
622 for giving warnings.
624 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
625 However, this was not working for (?!) because it is optimized to (*FAIL),
626 for which PCRE does not allow quantifiers. The optimization is now disabled
627 when a quantifier follows (?!). I can't see any use for this, but it makes
628 things uniform.
630 36. Perl no longer allows group names to start with digits, so I have made this
631 change also in PCRE. It simplifies the code a bit.
633 37. In extended mode, Perl ignores spaces before a + that indicates a
634 possessive quantifier. PCRE allowed a space before the quantifier, but not
635 before the possessive +. It now does.
637 38. The use of \K (reset reported match start) within a repeated possessive
638 group such as (a\Kb)*+ was not working.
640 40. Document that the same character tables must be used at compile time and
641 run time, and that the facility to pass tables to pcre_exec() and
642 pcre_dfa_exec() is for use only with saved/restored patterns.
644 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
645 features for Builds with MSVC:
647 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
648 builds). The .rc files can be used to set FileDescription and many other
649 attributes.
651 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
652 This allows higher-level build scripts which want .pdb files to avoid
653 hard-coding the exact files needed."
655 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
656 mean "start of word" and "end of word", respectively, as a transition aid.
658 43. A minimizing repeat of a class containing codepoints greater than 255 in
659 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
660 compiled to use the heap for recursion.
662 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
663 is configured.
666 Version 8.33 28-May-2013
667 ------------------------
669 1. Added 'U' to some constants that are compared to unsigned integers, to
670 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
671 variables that are added to signed variables, to ensure the result is
672 signed and can be negated.
674 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
675 CMake config files.
677 3. Revise the creation of config.h.generic so that all boolean macros are
678 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
679 overriding via -D on the command line possible.
681 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
682 to unsigned int is reported to make a quite noticeable speed difference in
683 a specific Windows environment. Testing on Linux did also appear to show
684 some benefit (and it is clearly not harmful). Also fixed the definition of
685 Xop which should be unsigned.
687 5. Related to (4), changing the definition of the intermediate variable cc
688 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
689 improvements.
691 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
692 unnecessary spaces.
694 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
695 and later.
697 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
699 9. Optimizing fast_forward_start_bits in JIT.
701 10. Adding support for callouts in JIT, and fixing some issues revealed
702 during this work. Namely:
704 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
706 (b) Minimum length was not checked before the matching is started.
708 11. The value of capture_last that is passed to callouts was incorrect in some
709 cases when there was a capture on one path that was subsequently abandoned
710 after a backtrack. Also, the capture_last value is now reset after a
711 recursion, since all captures are also reset in this case.
713 12. The interpreter no longer returns the "too many substrings" error in the
714 case when an overflowing capture is in a branch that is subsequently
715 abandoned after a backtrack.
717 13. In the pathological case when an offset vector of size 2 is used, pcretest
718 now prints out the matched string after a yield of 0 or 1.
720 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
721 Only supported by the JIT compiler at the moment.
723 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
725 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
726 offset of the starting point of the matching process, provided the offsets
727 vector is large enough.
729 17. The \A escape now records a lookbehind value of 1, though its execution
730 does not actually inspect the previous character. This is to ensure that,
731 in partial multi-segment matching, at least one character from the old
732 segment is retained when a new segment is processed. Otherwise, if there
733 are no lookbehinds in the pattern, \A might match incorrectly at the start
734 of a new segment.
736 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
738 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
739 gives some modest performance improvement in 8-bit mode.
741 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
742 be expressed in certain programming languages using Universal Character
743 Names.
745 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
746 which points out that "non characters" are not "characters that may not
747 appear in Unicode strings" but rather "characters that are reserved for
748 internal use and have only local meaning".
750 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
751 there was a conditional group that depended on an assertion, if the
752 assertion was false, the callout that immediately followed the alternation
753 in the condition was skipped when pcre_exec() was used for matching.
755 23. Allow an explicit callout to be inserted before an assertion that is the
756 condition for a conditional group, for compatibility with automatic
757 callouts, which always insert a callout at this point.
759 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
760 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
762 25. (*PRUNE) is now supported by the JIT compiler.
764 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
766 27. Fix the case where there are two or more SKIPs with arguments that may be
767 ignored.
769 28. (*SKIP) is now supported by the JIT compiler.
771 29. (*THEN) is now supported by the JIT compiler.
773 30. Update RunTest with additional test selector options.
775 31. The way PCRE handles backtracking verbs has been changed in two ways.
777 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
778 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
779 backtracking. In some cases this makes it more Perl-compatible, but Perl's
780 rather obscure rules do not always do the same thing.
782 (2) Previously, backtracking verbs were confined within assertions. This is
783 no longer the case for positive assertions, except for (*ACCEPT). Again,
784 this sometimes improves Perl compatibility, and sometimes does not.
786 32. A number of tests that were in test 2 because Perl did things differently
787 have been moved to test 1, because either Perl or PCRE has changed, and
788 these tests are now compatible.
790 32. Backtracking control verbs are now handled in the same way in JIT and
791 interpreter.
793 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
794 contained a forward subroutine reference caused a compile error.
796 34. Auto-detect and optimize limited repetitions in JIT.
798 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
799 blocking (*UTF) etc.
801 36. In the interpreter, maximizing pattern repetitions for characters and
802 character types now use tail recursion, which reduces stack usage.
804 37. The value of the max lookbehind was not correctly preserved if a compiled
805 and saved regex was reloaded on a host of different endianness.
807 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
808 of the compiled pattern block, expand the flags field from 16 to 32 bits
809 because it was almost full.
811 39. Try madvise first before posix_madvise.
813 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
814 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
815 original change locked out all matches of empty strings. This has been
816 changed so that one match of an empty string per line is recognized.
817 Subsequent searches on the same line (for colouring or for --only-matching,
818 for example) do not recognize empty strings.
820 41. Applied a user patch to fix a number of spelling mistakes in comments.
822 42. Data lines longer than 65536 caused pcretest to crash.
824 43. Clarified the data type for length and startoffset arguments for pcre_exec
825 and pcre_dfa_exec in the function-specific man pages, where they were
826 explicitly stated to be in bytes, never having been updated. I also added
827 some clarification to the pcreapi man page.
829 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
830 a segmentation fault.
833 Version 8.32 30-November-2012
834 -----------------------------
836 1. Improved JIT compiler optimizations for first character search and single
837 character iterators.
839 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
840 Patch by Daniel Richard G.
842 3. Single character iterator optimizations in the JIT compiler.
844 4. Improved JIT compiler optimizations for character ranges.
846 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
847 Reported by Giuseppe D'Angelo.
849 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
850 of a line, was being set incorrectly in cases where .* appeared inside
851 atomic brackets at the start of a pattern, or where there was a subsequent
852 *PRUNE or *SKIP.
854 7. Improved instruction cache flush for POWER/PowerPC.
855 Patch by Daniel Richard G.
857 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
858 grep:
860 (a) There is now no limit to the number of patterns to be matched.
862 (b) An error is given if a pattern is too long.
864 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
865 are now supported.
867 (d) --exclude-from and --include-from (multiple use) have been added.
869 (e) Exclusions and inclusions now apply to all files and directories, not
870 just to those obtained from scanning a directory recursively.
872 (f) Multiple uses of -f and --file-list are now supported.
874 (g) In a Windows environment, the default for -d has been changed from
875 "read" (the GNU grep default) to "skip", because otherwise the presence
876 of a directory in the file list provokes an error.
878 (h) The documentation has been revised and clarified in places.
880 9. Improve the matching speed of capturing brackets.
882 10. Changed the meaning of \X so that it now matches a Unicode extended
883 grapheme cluster.
885 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
886 out POSIX threads when JIT support is configured.
888 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
890 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
891 structure could go wrong in environments where size_t is not the same size
892 as int.
894 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
895 set.
897 15. The EBCDIC support had decayed; later updates to the code had included
898 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
899 general tidy up of EBCDIC-related issues, and the documentation was also
900 not quite right. There is now a test that can be run on ASCII systems to
901 check some of the EBCDIC-related things (but is it not a full test).
903 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
904 in a small tidy to the code.
906 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
908 18. If the --only-matching (-o) option in pcregrep is specified multiple
909 times, each one causes appropriate output. For example, -o1 -o2 outputs the
910 substrings matched by the 1st and 2nd capturing parentheses. A separating
911 string can be specified by --om-separator (default empty).
913 19. Improving the first n character searches.
915 20. Turn case lists for horizontal and vertical white space into macros so that
916 they are defined only once.
918 21. This set of changes together give more compatible Unicode case-folding
919 behaviour for characters that have more than one other case when UCP
920 support is available.
922 (a) The Unicode property table now has offsets into a new table of sets of
923 three or more characters that are case-equivalent. The MultiStage2.py
924 script that generates these tables (the pcre_ucd.c file) now scans
925 CaseFolding.txt instead of UnicodeData.txt for character case
926 information.
928 (b) The code for adding characters or ranges of characters to a character
929 class has been abstracted into a generalized function that also handles
930 case-independence. In UTF-mode with UCP support, this uses the new data
931 to handle characters with more than one other case.
933 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
934 whose other case is greater than 256 are now correctly matched
935 caselessly. Previously, the high codepoint matched the low one, but not
936 vice versa.
938 (d) The processing of \h, \H, \v, and \ in character classes now makes use
939 of the new class addition function, using character lists defined as
940 macros alongside the case definitions of 20 above.
942 (e) Caseless back references now work with characters that have more than
943 one other case.
945 (f) General caseless matching of characters with more than one other case
946 is supported.
948 22. Unicode character properties were updated from Unicode 6.2.0
950 23. Improved CMake support under Windows. Patch by Daniel Richard G.
952 24. Add support for 32-bit character strings, and UTF-32
954 25. Major JIT compiler update (code refactoring and bugfixing).
955 Experimental Sparc 32 support is added.
957 26. Applied a modified version of Daniel Richard G's patch to create
958 pcre.h.generic and config.h.generic by "make" instead of in the
959 PrepareRelease script.
961 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
962 pcre_compile.c when checking for a zero character.
964 28. Introducing a native interface for JIT. Through this interface, the compiled
965 machine code can be directly executed. The purpose of this interface is to
966 provide fast pattern matching, so several sanity checks are not performed.
967 However, feature tests are still performed. The new interface provides
968 1.4x speedup compared to the old one.
970 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
971 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
972 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
974 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
975 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
976 the "old" RFC 2279). Instead, it ended up passing a negative length to
977 pcre_exec().
979 31. Add support for GCC's visibility feature to hide internal functions.
981 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
982 "unknown -C option" after outputting 0 or 1.
984 33. There is now support for generating a code coverage report for the test
985 suite in environments where gcc is the compiler and lcov is installed. This
986 is mainly for the benefit of the developers.
988 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
989 unaddressable using valgrind annotations, allowing valgrind to detect
990 invalid memory accesses. This is mainly for the benefit of the developers.
992 25. (*UTF) can now be used to start a pattern in any of the three libraries.
994 26. Give configure error if --enable-cpp but no C++ compiler found.
997 Version 8.31 06-July-2012
998 -------------------------
1000 1. Fixing a wrong JIT test case and some compiler warnings.
1002 2. Removed a bashism from the RunTest script.
1004 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1005 to unsigned type, result still unsigned" that was given by an MS compiler
1006 on encountering the code "-sizeof(xxx)".
1008 4. Partial matching support is added to the JIT compiler.
1010 5. Fixed several bugs concerned with partial matching of items that consist
1011 of more than one character:
1013 (a) /^(..)\1/ did not partially match "aba" because checking references was
1014 done on an "all or nothing" basis. This also applied to repeated
1015 references.
1017 (b) \R did not give a hard partial match if \r was found at the end of the
1018 subject.
1020 (c) \X did not give a hard partial match after matching one or more
1021 characters at the end of the subject.
1023 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1024 a partial match for the string "\r".
1026 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1027 a partial match for a CR character at the end of the subject string.
1029 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1030 running pcretest, the text "(JIT)" added to the output whenever JIT is
1031 actually used to run the match.
1033 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1034 or /S+[+] with a digit between 1 and 7.
1036 8. OP_NOT now supports any UTF character not just single-byte ones.
1038 9. (*MARK) control verb is now supported by the JIT compiler.
1040 10. The command "./RunTest list" lists the available tests without actually
1041 running any of them. (Because I keep forgetting what they all are.)
1045 12. Applied a (slightly modified) user-supplied patch that improves performance
1046 when the heap is used for recursion (compiled with --disable-stack-for-
1047 recursion). Instead of malloc and free for each heap frame each time a
1048 logical recursion happens, frames are retained on a chain and re-used where
1049 possible. This sometimes gives as much as 30% improvement.
1051 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1052 call.
1054 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1056 15. It is now possible to link pcretest with libedit as an alternative to
1057 libreadline.
1059 16. (*COMMIT) control verb is now supported by the JIT compiler.
1061 17. The Unicode data tables have been updated to Unicode 6.1.0.
1063 18. Added --file-list option to pcregrep.
1065 19. Added binary file support to pcregrep, including the -a, --binary-files,
1066 -I, and --text options.
1068 20. The madvise function is renamed for posix_madvise for QNX compatibility
1069 reasons. Fixed by Giuseppe D'Angelo.
1071 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1073 22. Changed the option for creating symbolic links for 16-bit man pages from
1074 -s to -sf so that re-installing does not cause issues.
1076 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1078 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1079 matches in certain environments (the workspace was not being correctly
1080 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1081 some of the workspace data at the beginning of a restart.
1083 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1084 was not doing so when it should - probably a typo introduced by SVN 528
1085 (change 8.10/14).
1087 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1088 \w+ when the character tables indicated that \x{c4} was a word character.
1089 There were several related cases, all because the tests for doing a table
1090 lookup were testing for characters less than 127 instead of 255.
1092 27. If a pattern contains capturing parentheses that are not used in a match,
1093 their slots in the ovector are set to -1. For those that are higher than
1094 any matched groups, this happens at the end of processing. In the case when
1095 there were back references that the ovector was too small to contain
1096 (causing temporary malloc'd memory to be used during matching), and the
1097 highest capturing number was not used, memory off the end of the ovector
1098 was incorrectly being set to -1. (It was using the size of the temporary
1099 memory instead of the true size.)
1101 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1102 ovector size, it uses memory at the end of the block that it has got.
1104 29. Check for an overlong MARK name and give an error at compile time. The
1105 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1107 30. JIT compiler update.
1109 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1110 Rill for the patch.
1112 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1114 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1116 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1117 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1118 was enabled.
1120 35. Improve JIT code generation for greedy plus quantifier.
1122 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1123 1 to "aa" instead of to an empty string. The bug affected repeated groups
1124 that could potentially match an empty string.
1126 37. Optimizing single character iterators in JIT.
1128 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1129 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1130 codepoints that are too big for the mode are faulted, and in a UTF mode,
1131 disallowed codepoints are also faulted.
1133 39. If PCRE was compiled with UTF support, in three places in the DFA
1134 matcher there was code that should only have been obeyed in UTF mode, but
1135 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1136 processing when bytes with values greater than 127 were present. In 16-bit
1137 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1138 both cases the values are those that cannot be the first data item in a UTF
1139 character. The three items that might have provoked this were recursions,
1140 possessively repeated groups, and atomic groups.
1142 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1143 and pcregrep, because some OS require shared objects to be explicitly
1144 passed to ld, causing the link step to fail if they are not.
1146 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1147 mode, patterns that started with \h* or \R* might be incorrectly matched.
1150 Version 8.30 04-February-2012
1151 -----------------------------
1153 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1154 name is defined in ctype.h.
1156 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1157 only in quite long subpatterns.
1159 3. Removed the function pcre_info(), which has been obsolete and deprecated
1160 since it was replaced by pcre_fullinfo() in February 2000.
1162 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1163 match a (*MARK), and the match failed at the start of the subject, a
1164 reference to memory before the start of the subject could occur. This bug
1165 was introduced by fix 17 of release 8.21.
1167 5. A reference to an unset group with zero minimum repetition was giving
1168 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1169 /(another)?(\1?)test/ matched against "hello world test". This bug was
1170 introduced in release 8.13.
1172 6. Add support for 16-bit character strings (a large amount of work involving
1173 many changes and refactorings).
1175 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1176 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1177 from a file.
1179 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1180 rounding is not applied in this particular case).
1182 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1183 if they appear, or are escaped, in patterns.
1185 10. Get rid of a number of -Wunused-but-set-variable warnings.
1187 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1188 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1189 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1190 also returns the mark "x". This bug applied to capturing parentheses,
1191 non-capturing parentheses, and atomic parentheses. It also applied to some
1192 assertions.
1194 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1195 information out of configure.ac instead of relying on pcre.h.generic, which
1196 is not stored in the repository.
1198 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1199 -lreadline.
1201 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1203 15. Applied Graycode's patch to put the top-level frame on the stack rather
1204 than the heap when not using the stack for recursion. This gives a
1205 performance improvement in many cases when recursion is not deep.
1207 16. Experimental code added to "pcretest -C" to output the stack frame size.
1210 Version 8.21 12-Dec-2011
1211 ------------------------
1213 1. Updating the JIT compiler.
1215 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1216 are added as well.
1218 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1219 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1220 calling _pcre_jit_exec. Some extra comments are added.
1222 4. (*MARK) settings inside atomic groups that do not contain any capturing
1223 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1224 was introduced by change 18 for 8.20.
1226 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1227 ECMA-262 standard.
1229 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1230 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1231 This bug was probably introduced by change 9 of 8.13.
1233 7. While fixing 6 above, I noticed that a number of other items were being
1234 incorrectly rejected as "not fixed length". This arose partly because newer
1235 opcodes had not been added to the fixed-length checking code. I have (a)
1236 corrected the bug and added tests for these items, and (b) arranged for an
1237 error to occur if an unknown opcode is encountered while checking for fixed
1238 length instead of just assuming "not fixed length". The items that were
1239 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1240 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1241 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1243 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1244 being incorrectly compiled and would have given unpredicatble results.
1246 9. A possessively repeated subpattern with minimum repeat count greater than
1247 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1248 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1249 the first (A) could occur when it should not.
1251 10. Add a cast and remove a redundant test from the code.
1253 11. JIT should use pcre_malloc/pcre_free for allocation.
1255 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1256 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1257 is anything other than /usr, -L is still shown).
1259 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1261 14. Perl does not support \N without a following name in a [] class; PCRE now
1262 also gives an error.
1264 15. If a forward reference was repeated with an upper limit of around 2000,
1265 it caused the error "internal error: overran compiling workspace". The
1266 maximum number of forward references (including repeats) was limited by the
1267 internal workspace, and dependent on the LINK_SIZE. The code has been
1268 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1269 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1270 of around 200,000 forward references. While doing this, I also speeded up
1271 the filling in of repeated forward references.
1273 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1274 incorrectly expecting the subject to contain another "a" after the start.
1276 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1277 in the match, the SKIP should be ignored. This was not happening; instead
1278 the SKIP was being treated as NOMATCH. For patterns such as
1279 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1280 tested.
1282 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1283 now much more compatible with Perl, in particular in cases where the result
1284 is a non-match for a non-anchored pattern. For example, if
1285 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1286 "m", where previously it did not return a name. A side effect of this
1287 change is that for partial matches, the last encountered mark name is
1288 returned, as for non matches. A number of tests that were previously not
1289 Perl-compatible have been moved into the Perl-compatible test files. The
1290 refactoring has had the pleasing side effect of removing one argument from
1291 the match() function, thus reducing its stack requirements.
1293 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1294 subsequent uses of /S (without +) incorrectly behaved like /S+.
1296 21. Retrieve executable code size support for the JIT compiler and fixing
1297 some warnings.
1299 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1300 not work when the shorter character appeared right at the end of the
1301 subject string.
1303 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1304 systems.
1306 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1307 output it when the /M option is used in pcretest.
1309 25. The CheckMan script was not being included in the distribution. Also, added
1310 an explicit "perl" to run Perl scripts from the PrepareRelease script
1311 because this is reportedly needed in Windows.
1313 26. If study data was being save in a file and studying had not found a set of
1314 "starts with" bytes for the pattern, the data written to the file (though
1315 never used) was taken from uninitialized memory and so caused valgrind to
1316 complain.
1318 27. Updated RunTest.bat as provided by Sheri Pierce.
1320 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1322 29. Computation of memory usage for the table of capturing group names was
1323 giving an unnecessarily large value.
1326 Version 8.20 21-Oct-2011
1327 ------------------------
1329 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1330 a POSIX class. After further experiments with Perl, which convinced me that
1331 Perl has bugs and confusions, a closing square bracket is no longer allowed
1332 in a POSIX name. This bug also affected patterns with classes that started
1333 with full stops.
1335 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1336 captured substring, but while checking the failing first alternative,
1337 substring 1 is temporarily captured. If the output vector supplied to
1338 pcre_exec() was not big enough for this capture, the yield of the function
1339 was still zero ("insufficient space for captured substrings"). This cannot
1340 be totally fixed without adding another stack variable, which seems a lot
1341 of expense for a edge case. However, I have improved the situation in cases
1342 such as /(a)(b)x|abc/ matched against "abc", where the return code
1343 indicates that fewer than the maximum number of slots in the ovector have
1344 been set.
1346 3. Related to (2) above: when there are more back references in a pattern than
1347 slots in the output vector, pcre_exec() uses temporary memory during
1348 matching, and copies in the captures as far as possible afterwards. It was
1349 using the entire output vector, but this conflicts with the specification
1350 that only 2/3 is used for passing back captured substrings. Now it uses
1351 only the first 2/3, for compatibility. This is, of course, another edge
1352 case.
1354 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1355 main code base, and can be used by building with --enable-jit. When this is
1356 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1357 runtime --no-jit option is given.
1359 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1360 ovector, the return from the function was zero, implying that there were
1361 other matches that did not fit. The correct "exactly full" value is now
1362 returned.
1364 6. If a subpattern that was called recursively or as a subroutine contained
1365 (*PRUNE) or any other control that caused it to give a non-standard return,
1366 invalid errors such as "Error -26 (nested recursion at the same subject
1367 position)" or even infinite loops could occur.
1369 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1370 computing the minimum length on reaching *ACCEPT, and so ended up with the
1371 wrong value of 1 rather than 0. Further investigation indicates that
1372 computing a minimum subject length in the presence of *ACCEPT is difficult
1373 (think back references, subroutine calls), and so I have changed the code
1374 so that no minimum is registered for a pattern that contains *ACCEPT.
1376 8. If (*THEN) was present in the first (true) branch of a conditional group,
1377 it was not handled as intended. [But see 16 below.]
1379 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1380 Sheri Pierce.
1382 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1383 the first byte in a match must be "a".
1385 11. Change 17 for 8.13 increased the recursion depth for patterns like
1386 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1387 pattern contains any instances of (*THEN). If it does not, the old
1388 optimizations are restored. It would be nice to do this on a per-group
1389 basis, but at the moment that is not feasible.
1391 12. In some environments, the output of pcretest -C is CRLF terminated. This
1392 broke RunTest's code that checks for the link size. A single white space
1393 character after the value is now allowed for.
1395 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1396 For "fr", it uses the Windows-specific input and output files.
1398 14. If (*THEN) appeared in a group that was called recursively or as a
1399 subroutine, it did not work as intended. [But see next item.]
1401 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1402 pattern fragments (but not containing any | characters). If A and B are
1403 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1404 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1405 D. In other words, Perl considers parentheses that do not contain any |
1406 characters to be part of a surrounding alternative, whereas PCRE was
1407 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1408 differently. PCRE now behaves in the same way as Perl, except in the case
1409 of subroutine/recursion calls such as (?1) which have in any case always
1410 been different (but PCRE had them first :-).
1412 16. Related to 15 above: Perl does not treat the | in a conditional group as
1413 creating alternatives. Such a group is treated in the same way as an
1414 ordinary group without any | characters when processing (*THEN). PCRE has
1415 been changed to match Perl's behaviour.
1417 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1418 RunGrepTest script failed.
1420 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1421 inevitable for groups that contain captures, but it can lead to a lot of
1422 stack use in large patterns. The old behaviour has been restored for atomic
1423 groups that do not contain any capturing parentheses.
1425 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1426 suppress the check for a minimum subject length at run time. (If it was
1427 given to pcre_exec() or pcre_dfa_exec() it did work.)
1429 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1430 fail to work when decoding hex characters in data strings in EBCDIC
1431 environments.
1433 21. It appears that in at least one Mac OS environment, the isxdigit() function
1434 is implemented as a macro that evaluates to its argument more than once,
1435 contravening the C 90 Standard (I haven't checked a later standard). There
1436 was an instance in pcretest which caused it to go wrong when processing
1437 \x{...} escapes in subject strings. The has been rewritten to avoid using
1438 things like p++ in the argument of isxdigit().
1441 Version 8.13 16-Aug-2011
1442 ------------------------
1444 1. The Unicode data tables have been updated to Unicode 6.0.0.
1446 2. Two minor typos in pcre_internal.h have been fixed.
1448 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1449 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1450 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1452 4. There were a number of related bugs in the code for matching backrefences
1453 caselessly in UTF-8 mode when codes for the characters concerned were
1454 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1455 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1456 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1457 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1458 2-byte code at the end of the subject (it thought there wasn't enough data
1459 left).
1461 5. Comprehensive information about what went wrong is now returned by
1462 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1463 as the output vector has at least 2 elements. The offset of the start of
1464 the failing character and a reason code are placed in the vector.
1466 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1467 now returned is for the first byte of the failing character, instead of the
1468 last byte inspected. This is an incompatible change, but I hope it is small
1469 enough not to be a problem. It makes the returned offset consistent with
1470 pcre_exec() and pcre_dfa_exec().
1472 7. pcretest now gives a text phrase as well as the error number when
1473 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1474 failure, the offset and reason code are output.
1476 8. When \R was used with a maximizing quantifier it failed to skip backwards
1477 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1478 back over a single character (\n). This seems wrong (because it treated the
1479 two characters as a single entity when going forwards), conflicts with the
1480 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1481 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1482 has been changed.
1484 9. Some internal refactoring has changed the processing so that the handling
1485 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1486 time (the PCRE_DOTALL option was changed this way some time ago: version
1487 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1488 which was always a bit of a fudge. It also means that there is one less
1489 argument for the match() function, which reduces its stack requirements
1490 slightly. This change also fixes an incompatibility with Perl: the pattern
1491 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1493 10. More internal refactoring has drastically reduced the number of recursive
1494 calls to match() for possessively repeated groups such as (abc)++ when
1495 using pcre_exec().
1497 11. While implementing 10, a number of bugs in the handling of groups were
1498 discovered and fixed:
1500 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1501 (a|)*(?1) gave a compile-time internal error.
1502 ((a|)+)+ did not notice that the outer group could match an empty string.
1503 (^a|^)+ was not marked as anchored.
1504 (.*a|.*)+ was not marked as matching at start or after a newline.
1506 12. Yet more internal refactoring has removed another argument from the match()
1507 function. Special calls to this function are now indicated by setting a
1508 value in a variable in the "match data" data block.
1510 13. Be more explicit in pcre_study() instead of relying on "default" for
1511 opcodes that mean there is no starting character; this means that when new
1512 ones are added and accidentally left out of pcre_study(), testing should
1513 pick them up.
1515 14. The -s option of pcretest has been documented for ages as being an old
1516 synonym of -m (show memory usage). I have changed it to mean "force study
1517 for every regex", that is, assume /S for every regex. This is similar to -i
1518 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1519 using it. It makes it easier to run collections of tests with and without
1520 study enabled, and thereby test pcre_study() more easily. All the standard
1521 tests are now run with and without -s (but some patterns can be marked as
1522 "never study" - see 20 below).
1524 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1525 restoration of the capturing data to the outer values was not happening
1526 correctly.
1528 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1529 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1530 pattern had matched an empty string, and so incorrectly returned a no
1531 match.
1533 17. There was optimizing code for the last branch of non-capturing parentheses,
1534 and also for the obeyed branch of a conditional subexpression, which used
1535 tail recursion to cut down on stack usage. Unfortunately, now that there is
1536 the possibility of (*THEN) occurring in these branches, tail recursion is
1537 no longer possible because the return has to be checked for (*THEN). These
1538 two optimizations have therefore been removed. [But see 8.20/11 above.]
1540 18. If a pattern containing \R was studied, it was assumed that \R always
1541 matched two bytes, thus causing the minimum subject length to be
1542 incorrectly computed because \R can also match just one byte.
1544 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1545 was incorrectly computed.
1547 20. If /S is present twice on a test pattern in pcretest input, it now
1548 *disables* studying, thereby overriding the use of -s on the command line
1549 (see 14 above). This is necessary for one or two tests to keep the output
1550 identical in both cases.
1552 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1553 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1555 22. When an atomic group that contained a capturing parenthesis was
1556 successfully matched, but the branch in which it appeared failed, the
1557 capturing was not being forgotten if a higher numbered group was later
1558 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1559 group 1 to "a", when in fact it should be unset. This applied to multi-
1560 branched capturing and non-capturing groups, repeated or not, and also to
1561 positive assertions (capturing in negative assertions does not happen
1562 in PCRE) and also to nested atomic groups.
1564 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1565 subject after a captured substring, to make it easier to tell which of a
1566 number of identical substrings has been captured.
1568 24. The way atomic groups are processed by pcre_exec() has been changed so that
1569 if they are repeated, backtracking one repetition now resets captured
1570 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1571 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1572 "aaa". Previously, it would have been "a". As part of this code
1573 refactoring, the way recursive calls are handled has also been changed.
1575 25. If an assertion condition captured any substrings, they were not passed
1576 back unless some other capturing happened later. For example, if
1577 (?(?=(a))a) was matched against "a", no capturing was returned.
1579 26. When studying a pattern that contained subroutine calls or assertions,
1580 the code for finding the minimum length of a possible match was handling
1581 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1582 group 1 called group 2 while simultaneously a separate group 2 called group
1583 1). A stack overflow occurred in this case. I have fixed this by limiting
1584 the recursion depth to 10.
1586 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1587 Fortmann. This supports explicit test numbers on the command line, and has
1588 argument validation and error reporting.
1590 28. An instance of \X with an unlimited repeat could fail if at any point the
1591 first character it looked at was a mark character.
1593 29. Some minor code refactoring concerning Unicode properties and scripts
1594 should reduce the stack requirement of match() slightly.
1596 30. Added the '=' option to pcretest to check the setting of unused capturing
1597 slots at the end of the pattern, which are documented as being -1, but are
1598 not included in the return count.
1600 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1601 compiled something random. Now it gives a compile-time error (as does
1602 Perl).
1604 32. A *MARK encountered during the processing of a positive assertion is now
1605 recorded and passed back (compatible with Perl).
1607 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1608 had alternative anchored branches, the search for a second match in a line
1609 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1610 matched the line "0102" twice. The same bug affected patterns that started
1611 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1612 twice.
1614 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1615 does, and because of capturing effects, quantifying parenthesized
1616 assertions may at times be useful. Quantifiers are now allowed for
1617 parenthesized assertions.
1619 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1621 36. \g was being checked for fancy things in a character class, when it should
1622 just be a literal "g".
1624 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1625 appearance of a nested POSIX class supersedes an apparent external class.
1626 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1627 unescaped square brackets may also appear as part of class names. For
1628 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1629 more like Perl. (But see 8.20/1 above.)
1631 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1632 was because it thought it was \N{name}, which is not supported).
1634 39. Add minix to OS list not supporting the -S option in pcretest.
1636 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1637 cannot analyze patterns in sufficient detail to catch mutual recursions
1638 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1639 subgroup is called recursively as a subpattern for a second time at the
1640 same position in the subject string. In previous releases this might have
1641 been caught by the recursion limit, or it might have run out of stack.
1643 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1644 happen only once. PCRE was, however incorrectly giving a compile time error
1645 "recursive call could loop indefinitely" because it cannot analyze the
1646 pattern in sufficient detail. The compile time test no longer happens when
1647 PCRE is compiling a conditional subpattern, but actual runaway loops are
1648 now caught at runtime (see 40 above).
1650 42. It seems that Perl allows any characters other than a closing parenthesis
1651 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1652 has been changed to be the same.
1654 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1655 as not to get warnings when autogen.sh is called. Also changed
1656 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1658 44. To help people who use pcregrep to scan files containing exceedingly long
1659 lines, the following changes have been made:
1661 (a) The default value of the buffer size parameter has been increased from
1662 8K to 20K. (The actual buffer used is three times this size.)
1664 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1665 PCRE is built.
1667 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1668 to be set at run time.
1670 (d) Numerical values in pcregrep options can be followed by K or M, for
1671 example --buffer-size=50K.
1673 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1674 given and the return code is set to 2.
1676 45. Add a pointer to the latest mark to the callout data block.
1678 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1679 partial match of an empty string instead of no match. This was specific to
1680 the use of ".".
1682 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1683 complete match instead of a partial match. This bug was dependent on both
1684 the PCRE_UTF8 and PCRE_DOTALL options being set.
1686 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1687 starting byte set, because \b was not being ignored.
1690 Version 8.12 15-Jan-2011
1691 ------------------------
1693 1. Fixed some typos in the markup of the man pages, and wrote a script that
1694 checks for such things as part of the documentation building process.
1696 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1697 --match-limit and --recursion-limit options (added for 8.11). In
1698 particular, this made one of the standard tests fail. (The integer value
1699 went into the wrong half of a long int.)
1701 3. If the --colour option was given to pcregrep with -v (invert match), it
1702 did strange things, either producing crazy output, or crashing. It should,
1703 of course, ignore a request for colour when reporting lines that do not
1704 match.
1706 4. Another pcregrep bug caused similar problems if --colour was specified with
1707 -M (multiline) and the pattern match finished with a line ending.
1709 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1710 matched in multiline mode, the following line was shown as part of the
1711 match. This seems wrong, so I have changed it.
1713 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1714 the check for further matches in the same line (so they could be coloured)
1715 to overrun the end of the current line. If another match was found, it was
1716 incorrectly shown (and then shown again when found in the next line).
1718 7. If pcregrep was compiled under Windows, there was a reference to the
1719 function pcregrep_exit() before it was defined. I am assuming this was
1720 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1721 reported by a user. I've moved the definition above the reference.
1724 Version 8.11 10-Dec-2010
1725 ------------------------
1727 1. (*THEN) was not working properly if there were untried alternatives prior
1728 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1729 backtracked to try for "b" instead of moving to the next alternative branch
1730 at the same level (in this case, to look for "c"). The Perl documentation
1731 is clear that when (*THEN) is backtracked onto, it goes to the "next
1732 alternative in the innermost enclosing group".
1734 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1735 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1736 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1737 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1738 (*THEN).
1740 3. If \s appeared in a character class, it removed the VT character from
1741 the class, even if it had been included by some previous item, for example
1742 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1743 of \s, but is part of the POSIX "space" class.)
1745 4. A partial match never returns an empty string (because you can always
1746 match an empty string at the end of the subject); however the checking for
1747 an empty string was starting at the "start of match" point. This has been
1748 changed to the "earliest inspected character" point, because the returned
1749 data for a partial match starts at this character. This means that, for
1750 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1751 (previously it gave "no match").
1753 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1754 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1755 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1756 has an implication that the given string is incomplete (because a partial
1757 match is preferred over a full match). For this reason, these items now
1758 give a partial match in this situation. [Aside: previously, the one case
1759 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1760 match rather than a full match, which was wrong by the old rules, but is
1761 now correct.]
1763 6. There was a bug in the handling of #-introduced comments, recognized when
1764 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1765 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1766 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1767 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1768 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1769 places in pcre_compile().
1771 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1772 comments when looking ahead for named forward references to subpatterns,
1773 the only newline sequence it recognized was NL. It now handles newlines
1774 according to the set newline convention.
1776 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1777 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1778 cater for a lack of strerror(). These oversights have been fixed.
1780 9. Added --match-limit and --recursion-limit to pcregrep.
1782 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1784 11. When the -o option was used, pcregrep was setting a return code of 1, even
1785 when matches were found, and --line-buffered was not being honoured.
1787 12. Added an optional parentheses number to the -o and --only-matching options
1788 of pcregrep.
1790 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1791 can match an empty string. The code to do it in pcretest and pcredemo
1792 needed fixing:
1794 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1795 only one byte after an empty string match just before CRLF (this case
1796 just got forgotten; "any" and "anycrlf" were OK).
1798 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1799 mode when an empty string match preceded an ASCII character followed by
1800 a non-ASCII character. (The code for advancing by one character rather
1801 than one byte was nonsense.)
1803 (c) The pcredemo.c sample program did not have any code at all to handle
1804 the cases when CRLF is a valid newline sequence.
1806 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1807 as a starting offset was within the subject string. There is now a new
1808 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1809 negative or greater than the length of the string. In order to test this,
1810 pcretest is extended to allow the setting of negative starting offsets.
1812 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1813 starting offset points to the beginning of a UTF-8 character was
1814 unnecessarily clumsy. I tidied it up.
1816 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1817 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1819 17. Nobody had reported that the --include_dir option, which was added in
1820 release 7.7 should have been called --include-dir (hyphen, not underscore)
1821 for compatibility with GNU grep. I have changed it to --include-dir, but
1822 left --include_dir as an undocumented synonym, and the same for
1823 --exclude-dir, though that is not available in GNU grep, at least as of
1824 release 2.5.4.
1826 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1827 characters from a string of bytes) have been redefined so as not to use
1828 loops, in order to improve performance in some environments. At the same
1829 time, I abstracted some of the common code into auxiliary macros to save
1830 repetition (this should not affect the compiled code).
1832 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1833 compile-time error is now given if \c is not followed by an ASCII
1834 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1835 different, and any byte value is allowed.)
1837 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1838 START_OPTIMIZE option, which is now allowed at compile time - but just
1839 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1840 to pcregrep and other applications that have no direct access to PCRE
1841 options. The new /Y option in pcretest sets this option when calling
1842 pcre_compile().
1844 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1845 back references. Groups containing recursive back references were forced to
1846 be atomic by that change, but in the case of named groups, the amount of
1847 memory required was incorrectly computed, leading to "Failed: internal
1848 error: code overflow". This has been fixed.
1850 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1851 pcretest.c, to avoid build problems in some Borland environments.
1854 Version 8.10 25-Jun-2010
1855 ------------------------
1857 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1858 THEN.
1860 2. (*ACCEPT) was not working when inside an atomic group.
1862 3. Inside a character class, \B is treated as a literal by default, but
1863 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1864 causes the error). The code is unchanged, but I tidied the documentation.
1866 4. Inside a character class, PCRE always treated \R and \X as literals,
1867 whereas Perl faults them if its -w option is set. I have changed PCRE so
1868 that it faults them when PCRE_EXTRA is set.
1870 5. Added support for \N, which always matches any character other than
1871 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1873 6. When compiling pcregrep with newer versions of gcc which may have
1874 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1875 declared with attribute warn_unused_result" were given. Just casting the
1876 result to (void) does not stop the warnings; a more elaborate fudge is
1877 needed. I've used a macro to implement this.
1879 7. Minor change to pcretest.c to avoid a compiler warning.
1881 8. Added four artifical Unicode properties to help with an option to make
1882 \s etc use properties (see next item). The new properties are: Xan
1883 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1885 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1886 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1887 this option. Modified pcretest to add /W to test this facility. Added
1888 REG_UCP to make it available via the POSIX interface.
1890 10. Added --line-buffered to pcregrep.
1892 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1893 studied, and the match started with a letter with a code point greater than
1894 127 whose first byte was different to the first byte of the other case of
1895 the letter, the other case of this starting letter was not recognized
1896 (#976).
1898 12. If a pattern that was studied started with a repeated Unicode property
1899 test, for example, \p{Nd}+, there was the theoretical possibility of
1900 setting up an incorrect bitmap of starting bytes, but fortunately it could
1901 not have actually happened in practice until change 8 above was made (it
1902 added property types that matched character-matching opcodes).
1904 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1905 possible starting bytes for non-anchored patterns.
1907 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1908 \R, and also a number of cases that involve Unicode properties, both
1909 explicit and implicit when PCRE_UCP is set.
1911 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1912 input, it could crash or give wrong results if characters with values
1913 greater than 0xc0 were present in the subject string. (Detail: it assumed
1914 UTF-8 input when processing these items.)
1916 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1917 size_t is 64-bit (#991).
1919 17. Added a check for running out of memory when PCRE is compiled with
1920 --disable-stack-for-recursion (#990).
1922 18. If the last data line in a file for pcretest does not have a newline on
1923 the end, a newline was missing in the output.
1925 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1926 less than 128) in its various bitmaps. However, there is a facility for
1927 generating tables according to the current locale when PCRE is compiled. It
1928 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1929 characters, are recognized by isspace() and therefore were getting set in
1930 these tables, and indeed these tables seem to approximate to ISO 8859. This
1931 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1932 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1933 which of course cannot start UTF-8 characters. I have changed the code so
1934 that only real ASCII characters (less than 128) and the correct starting
1935 bytes for UTF-8 encodings are set for characters greater than 127 when in
1936 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1937 altogether.)
1939 20. Added the /T option to pcretest so as to be able to run tests with non-
1940 standard character tables, thus making it possible to include the tests
1941 used for 19 above in the standard set of tests.
1943 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1944 reference to a subpattern the other side of a comment that contains an
1945 opening parenthesis caused either an internal compiling error, or a
1946 reference to the wrong subpattern.
1949 Version 8.02 19-Mar-2010
1950 ------------------------
1952 1. The Unicode data tables have been updated to Unicode 5.2.0.
1954 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1955 configured.
1957 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1958 original author of that file, following a query about its status.
1960 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1961 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1963 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1964 quantifier applied to a forward-referencing subroutine call, could compile
1965 incorrect code or give the error "internal error: previously-checked
1966 referenced subpattern not found".
1968 6. Both MS Visual Studio and Symbian OS have problems with initializing
1969 variables to point to external functions. For these systems, therefore,
1970 pcre_malloc etc. are now initialized to local functions that call the
1971 relevant global functions.
1973 7. There were two entries missing in the vectors called coptable and poptable
1974 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1975 I've fixed the data, and added a kludgy way of testing at compile time that
1976 the lengths are correct (equal to the number of opcodes).
1978 8. Following on from 7, I added a similar kludge to check the length of the
1979 eint vector in pcreposix.c.
1981 9. Error texts for pcre_compile() are held as one long string to avoid too
1982 much relocation at load time. To find a text, the string is searched,
1983 counting zeros. There was no check for running off the end of the string,
1984 which could happen if a new error number was added without updating the
1985 string.
1987 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1989 11. \K was not working if it appeared in an atomic group or in a group that
1990 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1991 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1992 the assertion is positive, but not if it is negative.
1994 12. Change 11 fortuitously reduced the size of the stack frame used in the
1995 "match()" function of pcre_exec.c by one pointer. Forthcoming
1996 implementation of support for (*MARK) will need an extra pointer on the
1997 stack; I have reserved it now, so that the stack frame size does not
1998 decrease.
2000 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2001 item in branch that calls a recursion is a subroutine call - as in the
2002 second branch in the above example - was incorrectly given the compile-
2003 time error "recursive call could loop indefinitely" because pcre_compile()
2004 was not correctly checking the subroutine for matching a non-empty string.
2006 14. The checks for overrunning compiling workspace could trigger after an
2007 overrun had occurred. This is a "should never occur" error, but it can be
2008 triggered by pathological patterns such as hundreds of nested parentheses.
2009 The checks now trigger 100 bytes before the end of the workspace.
2011 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2014 Version 8.01 19-Jan-2010
2015 ------------------------
2017 1. If a pattern contained a conditional subpattern with only one branch (in
2018 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2019 computed the wrong minimum data length (which is of course zero for such
2020 subpatterns). This could cause incorrect "no match" results.
2022 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2023 the pattern is reset in the first branch, pcre_compile() failed with
2024 "internal error: code overflow at offset...". This happened only when
2025 the reset was to the original external option setting. (An optimization
2026 abstracts leading options settings into an external setting, which was the
2027 cause of this.)
2029 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2030 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2031 assertion pattern did not match (meaning that the assertion was true), it
2032 was incorrectly treated as false if the SKIP had been reached during the
2033 matching. This also applied to assertions used as conditions.
2035 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2036 assertion subpattern, including such a pattern used as a condition,
2037 unpredictable results occurred, instead of the error return
2040 5. The C++ GlobalReplace function was not working like Perl for the special
2041 situation when an empty string is matched. It now does the fancy magic
2042 stuff that is necessary.
2044 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2045 removed. (These were left over from very, very early versions of PCRE.)
2047 7. Some cosmetic changes to the code to make life easier when compiling it
2048 as part of something else:
2050 (a) Change DEBUG to PCRE_DEBUG.
2052 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2053 called "current" as "current_branch", to prevent a collision with the
2054 Linux macro when compiled as a kernel module.
2056 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2057 prevent a collision with the Linux macro when compiled as a kernel
2058 module.
2060 8. In pcre_compile() there are some checks for integer overflows that used to
2061 cast potentially large values to (double). This has been changed to that
2062 when building, a check for int64_t is made, and if it is found, it is used
2063 instead, thus avoiding the use of floating point arithmetic. (There is no
2064 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2065 double.
2067 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2068 2005 (difference between two addresses compared to an unsigned value).
2070 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2071 custom one, because of the following reported problem in Windows:
2073 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2074 under Win32.
2075 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2076 therefore missing the function definition.
2077 - The compiler thus generates a "C" signature for the test function.
2078 - The linker fails to find the "C" function.
2079 - PCRE fails to configure if asked to do so against libbz2.
2081 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2082 messages were output:
2084 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2085 rerunning libtoolize, to keep the correct libtool macros in-tree.
2086 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2088 I have done both of these things.
2090 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2091 most of the time, it *can* run out if it is given a pattern that contains a
2092 runaway infinite recursion. I updated the discussion in the pcrestack man
2093 page.
2095 13. Now that we have gone to the x.xx style of version numbers, the minor
2096 version may start with zero. Using 08 or 09 is a bad idea because users
2097 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2098 interpreted as invalid octal numbers. I've updated the previous comment in
2099 configure.ac, and also added a check that gives an error if 08 or 09 are
2100 used.
2102 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2103 causing partial matching to fail when the end of the subject matched \W
2104 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2106 15. There were some discrepancies between the declarations in pcre_internal.h
2107 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2108 their definitions. The declarations used "const uschar *" and the
2109 definitions used USPTR. Even though USPTR is normally defined as "const
2110 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2111 reported that: "This difference in casting confuses some C++ compilers, for
2112 example, SunCC recognizes above declarations as different functions and
2113 generates broken code for hbpcre." I have changed the declarations to use
2114 USPTR.
2116 16. GNU libtool is named differently on some systems. The autogen.sh script now
2117 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2118 (FreeBSD).
2120 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2121 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2122 comment: "Figure out how to create a longlong from a string: strtoll and
2123 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2124 instance, but it only takes 2 args instead of 3!"
2126 18. A subtle bug concerned with back references has been fixed by a change of
2127 specification, with a corresponding code fix. A pattern such as
2128 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2129 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2130 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2131 same bug. Such groups have to be quantified to be useful, or contained
2132 inside another quantified group. (If there's no repetition, the reference
2133 can never match.) The problem arises because, having left the group and
2134 moved on to the rest of the pattern, a later failure that backtracks into
2135 the group uses the captured value from the final iteration of the group
2136 rather than the correct earlier one. I have fixed this in PCRE by forcing
2137 any group that contains a reference to itself to be an atomic group; that
2138 is, there cannot be any backtracking into it once it has completed. This is
2139 similar to recursive and subroutine calls.
2142 Version 8.00 19-Oct-09
2143 ----------------------
2145 1. The table for translating pcre_compile() error codes into POSIX error codes
2146 was out-of-date, and there was no check on the pcre_compile() error code
2147 being within the table. This could lead to an OK return being given in
2148 error.
2150 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2151 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2152 in a Windows environment.
2154 3. The pcregrep --count option prints the count for each file even when it is
2155 zero, as does GNU grep. However, pcregrep was also printing all files when
2156 --files-with-matches was added. Now, when both options are given, it prints
2157 counts only for those files that have at least one match. (GNU grep just
2158 prints the file name in this circumstance, but including the count seems
2159 more useful - otherwise, why use --count?) Also ensured that the
2160 combination -clh just lists non-zero counts, with no names.
2162 4. The long form of the pcregrep -F option was incorrectly implemented as
2163 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2164 but it seems right to fix it, and I didn't think it was worth preserving
2165 the old behaviour.
2167 5. The command line items --regex=pattern and --regexp=pattern were not
2168 recognized by pcregrep, which required --regex pattern or --regexp pattern
2169 (with a space rather than an '='). The man page documented the '=' forms,
2170 which are compatible with GNU grep; these now work.
2172 6. No libpcreposix.pc file was created for pkg-config; there was just
2173 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2175 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2176 when UCP support is not needed, by modifying the Python script that
2177 generates it from Unicode data files. This should not matter if the module
2178 is correctly used as a library, but I received one complaint about 50K of
2179 unwanted data. My guess is that the person linked everything into his
2180 program rather than using a library. Anyway, it does no harm.
2182 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2183 was a minimum greater than 1 for a wide character in a possessive
2184 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2185 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2186 character. Chaos in the form of incorrect output or a compiling loop could
2187 result.
2189 9. The restrictions on what a pattern can contain when partial matching is
2190 requested for pcre_exec() have been removed. All patterns can now be
2191 partially matched by this function. In addition, if there are at least two
2192 slots in the offset vector, the offset of the earliest inspected character
2193 for the match and the offset of the end of the subject are set in them when
2194 PCRE_ERROR_PARTIAL is returned.
2196 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2197 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2198 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2199 and may be more useful for multi-segment matching.
2201 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2202 used to be given if ever the end of the subject was reached; now it is
2203 given only if matching could not proceed because another character was
2204 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2205 string "Z", which now yields "no match" instead of "partial match". In the
2206 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2207 final character ended with (*FAIL).
2209 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2210 if the pattern had a "must contain" character that was already found in the
2211 earlier partial match, unless partial matching was again requested. For
2212 example, with the pattern /dog.(body)?/, the "must contain" character is
2213 "g". If the first part-match was for the string "dog", restarting with
2214 "sbody" failed. This bug has been fixed.
2216 13. The string returned by pcre_dfa_exec() after a partial match has been
2217 changed so that it starts at the first inspected character rather than the
2218 first character of the match. This makes a difference only if the pattern
2219 starts with a lookbehind assertion or \b or \B (\K is not supported by
2220 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2221 matching functions compatible, and I think it's the right thing to do.
2223 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2224 so that the demonstration program is easily available in environments where
2225 PCRE has not been installed from source.
2227 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2228 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2229 library.
2231 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2232 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2233 is not the first non-POSIX option to be added. Clearly some people find
2234 these options useful.
2236 17. If a caller to the POSIX matching function regexec() passes a non-zero
2237 value for nmatch with a NULL value for pmatch, the value of
2238 nmatch is forced to zero.
2240 18. RunGrepTest did not have a test for the availability of the -u option of
2241 the diff command, as RunTest does. It now checks in the same way as
2242 RunTest, and also checks for the -b option.
2244 19. If an odd number of negated classes containing just a single character
2245 interposed, within parentheses, between a forward reference to a named
2246 subpattern and the definition of the subpattern, compilation crashed with
2247 an internal error, complaining that it could not find the referenced
2248 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2249 [The bug was that it was starting one character too far in when skipping
2250 over the character class, thus treating the ] as data rather than
2251 terminating the class. This meant it could skip too much.]
2253 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2254 /g option in pcretest when the pattern contains \K, which makes it possible
2255 to have an empty string match not at the start, even when the pattern is
2256 anchored. Updated pcretest and pcredemo to use this option.
2258 21. If the maximum number of capturing subpatterns in a recursion was greater
2259 than the maximum at the outer level, the higher number was returned, but
2260 with unset values at the outer level. The correct (outer level) value is
2261 now given.
2263 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2264 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2265 make it do so. The string so far is captured, making this feature
2266 compatible with Perl.
2268 23. The tests have been re-organized, adding tests 11 and 12, to make it
2269 possible to check the Perl 5.10 features against Perl 5.10.
2271 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2272 pattern matches a fixed length string. PCRE did not allow this; now it
2273 does. Neither allows recursion.
2275 25. I finally figured out how to implement a request to provide the minimum
2276 length of subject string that was needed in order to match a given pattern.
2277 (It was back references and recursion that I had previously got hung up
2278 on.) This code has now been added to pcre_study(); it finds a lower bound
2279 to the length of subject needed. It is not necessarily the greatest lower
2280 bound, but using it to avoid searching strings that are too short does give
2281 some useful speed-ups. The value is available to calling programs via
2282 pcre_fullinfo().
2284 26. While implementing 25, I discovered to my embarrassment that pcretest had
2285 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2286 study optimizations had never been tested with that matching function.
2287 Oops. What is worse, even when it was passed study data, there was a bug in
2288 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2289 were also very few tests of studied patterns with pcre_dfa_exec().
2291 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2292 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2293 on the other side of the coin, they are no longer allowed to have different
2294 names, because these cannot be distinguished in PCRE, and this has caused
2295 confusion. (This is a difference from Perl.)
2297 28. When duplicate subpattern names are present (necessarily with different
2298 numbers, as required by 27 above), and a test is made by name in a
2299 conditional pattern, either for a subpattern having been matched, or for
2300 recursion in such a pattern, all the associated numbered subpatterns are
2301 tested, and the overall condition is true if the condition is true for any
2302 one of them. This is the way Perl works, and is also more like the way
2303 testing by number works.
2306 Version 7.9 11-Apr-09
2307 ---------------------
2309 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2310 (pcretest), all targets were linked against these libraries. This included
2311 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2312 libraries. This caused unwanted dependencies to be created. This problem
2313 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2314 pcretest is linked with readline.
2316 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2317 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2318 moved outside it again, because FALSE and TRUE are already defined in AIX,
2319 but BOOL is not.
2321 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2322 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2324 4. The pcregrep documentation said spaces were inserted as well as colons (or
2325 hyphens) following file names and line numbers when outputting matching
2326 lines. This is not true; no spaces are inserted. I have also clarified the
2327 wording for the --colour (or --color) option.
2329 5. In pcregrep, when --colour was used with -o, the list of matching strings
2330 was not coloured; this is different to GNU grep, so I have changed it to be
2331 the same.
2333 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2334 each matching line was coloured. Now it goes on to look for further matches
2335 of any of the test patterns, which is the same behaviour as GNU grep.
2337 7. A pattern that could match an empty string could cause pcregrep to loop; it
2338 doesn't make sense to accept an empty string match in pcregrep, so I have
2339 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2340 seems to be how GNU grep behaves. [But see later change 40 for release
2341 8.33.]
2343 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2344 start or after a newline", because the conditional assertion was not being
2345 correctly handled. The rule now is that both the assertion and what follows
2346 in the first alternative must satisfy the test.
2348 9. If auto-callout was enabled in a pattern with a conditional group whose
2349 condition was an assertion, PCRE could crash during matching, both with
2350 pcre_exec() and pcre_dfa_exec().
2352 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2353 used for matching.
2355 11. Unicode property support in character classes was not working for
2356 characters (bytes) greater than 127 when not in UTF-8 mode.
2358 12. Added the -M command line option to pcretest.
2360 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2362 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2364 16. Added comments and documentation about mis-use of no_arg in the C++
2365 wrapper.
2367 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2368 from Martin Jerabek that uses macro names for all relevant character and
2369 string constants.
2371 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2372 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2373 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2374 these, but not everybody uses configure.
2376 19. A conditional group that had only one branch was not being correctly
2377 recognized as an item that could match an empty string. This meant that an
2378 enclosing group might also not be so recognized, causing infinite looping
2379 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2380 with the subject "ab", where knowledge that the repeated group can match
2381 nothing is needed in order to break the loop.
2383 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2384 exec(), but without supplying a callout function, matching went wrong.
2386 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2387 leak if the size of the offset vector was greater than 30. When the vector
2388 is smaller, the saved offsets during recursion go onto a local stack
2389 vector, but for larger vectors malloc() is used. It was failing to free
2390 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2391 error, in fact).
2393 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2394 heapframe that is used only when UTF-8 support is enabled. This caused no
2395 problem, but was untidy.
2397 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2398 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2399 included within another project.
2401 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2402 slightly modified by me:
2404 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2405 not building pcregrep.
2407 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2408 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2410 25. Forward references, both numeric and by name, in patterns that made use of
2411 duplicate group numbers, could behave incorrectly or give incorrect errors,
2412 because when scanning forward to find the reference group, PCRE was not
2413 taking into account the duplicate group numbers. A pattern such as
2414 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2416 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2417 the feature of a custom pointer more persuasive (as requested by a user).
2419 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2420 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2421 pre-defined.
2423 28. Added support for (*UTF8) at the start of a pattern.
2425 29. Arrange for flags added by the "release type" setting in CMake to be shown
2426 in the configuration summary.
2429 Version 7.8 05-Sep-08
2430 ---------------------
2432 1. Replaced UCP searching code with optimized version as implemented for Ad
2433 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2434 stage table and inline lookup instead of a function, giving speed ups of 2
2435 to 5 times on some simple patterns that I tested. Permission was given to
2436 distribute the MultiStage2.py script that generates the tables (it's not in
2437 the tarball, but is in the Subversion repository).
2439 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2440 scripts.
2442 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2443 a group with a zero qualifier. The result of the study could be incorrect,
2444 or the function might crash, depending on the pattern.
2446 4. Caseless matching was not working for non-ASCII characters in back
2447 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2448 It now works when Unicode Property Support is available.
2450 5. In pcretest, an escape such as \x{de} in the data was always generating
2451 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2452 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2453 truncation.
2455 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2457 7. Added two (int) casts to pcregrep when printing the difference of two
2458 pointers, in case they are 64-bit values.
2460 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2461 test 2 if it fails.
2463 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2464 and a #define of that name to empty if it is not externally set. This is to
2465 allow users of MSVC to set it if necessary.
2467 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2468 the convenience functions in the pcre_get.c source file.
2470 11. An option change at the start of a pattern that had top-level alternatives
2471 could cause overwriting and/or a crash. This command provoked a crash in
2472 some environments:
2474 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2476 This potential security problem was recorded as CVE-2008-2371.
2478 12. For a pattern where the match had to start at the beginning or immediately
2479 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2480 pcre_dfa_exec() could read past the end of the passed subject if there was
2481 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2482 pcretest so that it places the subject at the end of its malloc-ed buffer.
2484 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2485 exec() might read past the end of the data buffer in UTF-8 mode.
2487 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2488 the data contained the byte 0x85 as part of a UTF-8 character within its
2489 first line. This applied both to normal and DFA matching.
2491 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2492 /^[^d]*?$/8 failed to match "abc".
2494 16. Added a missing copyright notice to pcrecpp_internal.h.
2496 17. Make it more clear in the documentation that values returned from
2497 pcre_exec() in ovector are byte offsets, not character counts.
2499 18. Tidied a few places to stop certain compilers from issuing warnings.
2501 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2502 supplied by Stefan Weber. I made a further small update for 7.8 because
2503 there is a change of source arrangements: the pcre_searchfuncs.c module is
2504 replaced by pcre_ucd.c.
2507 Version 7.7 07-May-08
2508 ---------------------
2510 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2511 a string to a long long, pretend we don't even have a long long." This is
2512 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2514 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2515 pre-7.6 versions, which defined a global no_arg variable instead of putting
2516 it in the RE class. (See also #8 below.)
2518 3. Remove a line of dead code, identified by coverity and reported by Nuno
2519 Lopes.
2521 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2523 (1) The include/exclude patterns were being applied to the whole pathnames
2524 of files, instead of just to the final components.
2526 (2) If there was more than one level of directory, the subdirectories were
2527 skipped unless they satisfied the include/exclude conditions. This is
2528 inconsistent with GNU grep (and could even be seen as contrary to the
2529 pcregrep specification - which I improved to make it absolutely clear).
2530 The action now is always to scan all levels of directory, and just
2531 apply the include/exclude patterns to regular files.
2533 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2534 --exclude_dir in the tests to avoid scanning .svn directories.
2536 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2537 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2538 doesn't support NULs in patterns.
2540 7. Added some missing "const"s to declarations of static tables in
2541 pcre_compile.c and pcre_dfa_exec.c.
2543 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2544 caused by fix #2 above. (Subsequently also a second patch to fix the
2545 first patch. And a third patch - this was a messy problem.)
2547 9. Applied Craig's patch to remove the use of push_back().
2549 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2550 matching function regexec().
2552 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2553 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2554 references. PCRE supports relative numbers with this syntax (I don't think
2555 Oniguruma does).
2557 12. Previously, a group with a zero repeat such as (...){0} was completely
2558 omitted from the compiled regex. However, this means that if the group
2559 was called as a subroutine from elsewhere in the pattern, things went wrong
2560 (an internal error was given). Such groups are now left in the compiled
2561 pattern, with a new opcode that causes them to be skipped at execution
2562 time.
2564 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2565 to the way PCRE behaves:
2567 (a) A lone ] character is dis-allowed (Perl treats it as data).
2569 (b) A back reference to an unmatched subpattern matches an empty string
2570 (Perl fails the current match path).
2572 (c) A data ] in a character class must be notated as \] because if the
2573 first data character in a class is ], it defines an empty class. (In
2574 Perl it is not possible to have an empty class.) The empty class []
2575 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2576 The negative empty class [^] matches any one character, independently
2577 of the DOTALL setting.
2579 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2580 non-existent subpattern following a character class starting with ']' and
2581 containing () gave an internal compiling error instead of "reference to
2582 non-existent subpattern". Fortunately, when the pattern did exist, the
2583 compiled code was correct. (When scanning forwards to check for the
2584 existence of the subpattern, it was treating the data ']' as terminating
2585 the class, so got the count wrong. When actually compiling, the reference
2586 was subsequently set up correctly.)
2588 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2589 it was being rejected as not supported by pcre_dfa_exec(), even though
2590 other assertions are supported. I have made pcre_dfa_exec() support
2591 (*FAIL).
2593 16. The implementation of 13c above involved the invention of a new opcode,
2594 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2595 cannot be changed at match time, I realized I could make a small
2596 improvement to matching performance by compiling OP_ALLANY instead of
2597 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2598 on the OP_ANY path.
2600 17. Compiling pcretest on Windows with readline support failed without the
2601 following two fixes: (1) Make the unistd.h include conditional on
2602 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2604 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2605 ncurses library to be included for pcretest when ReadLine support is
2606 requested, but also to allow for it to be overridden. This patch came from
2607 Daniel Bergström.
2609 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2610 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2611 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2612 spotting this.
2615 Version 7.6 28-Jan-08
2616 ---------------------
2618 1. A character class containing a very large number of characters with
2619 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2620 overflow.
2622 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2623 HAVE_LONG_LONG is not defined.
2625 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2626 bring them up to date and include new features. This patch includes:
2628 - Fixed PH's badly added libz and libbz2 support.
2629 - Fixed a problem with static linking.
2630 - Added pcredemo. [But later removed - see 7 below.]
2631 - Fixed dftables problem and added an option.
2632 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2634 - Added readline support for pcretest.
2635 - Added an listing of the option settings after cmake has run.
2637 4. A user submitted a patch to Makefile that makes it easy to create
2638 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2639 Makefile.am that cause it to include this special target, without
2640 affecting anything else. Note that the same mingw target plus all
2641 the other distribution libraries and programs are now supported
2642 when configuring with CMake (see 6 below) instead of with
2643 Configure/Make.
2645 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2646 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2647 exported in the Windows port". It has not yet been confirmed that the patch
2648 solves the problem, but it does no harm.
2650 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2651 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2652 with CMake, and also correct the comment about stack recursion.
2654 7. Remove the automatic building of pcredemo from the ./configure system and
2655 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2656 of a program that users should build themselves after PCRE is installed, so
2657 building it automatically is not really right. What is more, it gave
2658 trouble in some build environments.
2660 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2663 Version 7.5 10-Jan-08
2664 ---------------------
2666 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2667 values in parens when parsing an RE using the C++ wrapper."
2669 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2670 Characters greater than 255 were excluded from the class instead of being
2671 included.
2673 3. The same bug as (2) above applied to negated POSIX classes such as
2674 [:^space:].
2676 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2677 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2678 I have changed it.
2680 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2681 first named subpattern) and a construct such as (?&a) would reference the
2682 first named subpattern whose name started with "a" (in other words, the
2683 length check was missing). Both these problems are fixed. "Subpattern name
2684 expected" is now given for (?&) (a zero-length name), and this patch also
2685 makes it give the same error for \k'' (previously it complained that that
2686 was a reference to a non-existent subpattern).
2688 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2689 this is right because (?- can be followed by option settings as well as by
2690 digits. I have, however, made the messages clearer.
2692 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2693 than the number used in the conditional) now cause a compile-time error.
2694 This is actually not compatible with Perl, which accepts such patterns, but
2695 treats the conditional as always being FALSE (as PCRE used to), but it
2696 seems to me that giving a diagnostic is better.
2698 8. Change "alphameric" to the more common word "alphanumeric" in comments
2699 and messages.
2701 9. Fix two occurrences of "backslash" in comments that should have been
2702 "backspace".
2704 10. Remove two redundant lines of code that can never be obeyed (their function
2705 was moved elsewhere).
2707 11. The program that makes PCRE's Unicode character property table had a bug
2708 which caused it to generate incorrect table entries for sequences of
2709 characters that have the same character type, but are in different scripts.
2710 It amalgamated them into a single range, with the script of the first of
2711 them. In other words, some characters were in the wrong script. There were
2712 thirteen such cases, affecting characters in the following ranges:
2714 U+002b0 - U+002c1
2715 U+0060c - U+0060d
2716 U+0061e - U+00612
2717 U+0064b - U+0065e
2718 U+0074d - U+0076d
2719 U+01800 - U+01805
2720 U+01d00 - U+01d77
2721 U+01d9b - U+01dbf
2722 U+0200b - U+0200f
2723 U+030fc - U+030fe
2724 U+03260 - U+0327f
2725 U+0fb46 - U+0fbb1
2726 U+10450 - U+1049d
2728 12. The -o option (show only the matching part of a line) for pcregrep was not
2729 compatible with GNU grep in that, if there was more than one match in a
2730 line, it showed only the first of them. It now behaves in the same way as
2731 GNU grep.
2733 13. If the -o and -v options were combined for pcregrep, it printed a blank
2734 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2735 does the same. The return code can be used to tell if there were any
2736 non-matching lines.
2738 14. Added --file-offsets and --line-offsets to pcregrep.
2740 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2741 infinitely looping recursion. The bug was that positive lookaheads were not
2742 being skipped when checking for a possible empty match (negative lookaheads
2743 and both kinds of lookbehind were skipped).
2745 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2746 inclusion of <windows.h> to before rather than after the definition of
2747 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2749 17. Specifying a possessive quantifier with a specific limit for a Unicode
2750 character property caused pcre_compile() to compile bad code, which led at
2751 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2752 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2753 caused the error; without that there was no problem.
2755 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2757 19. Added --enable-pcretest-libreadline.
2759 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2760 RE::GlobalReplace(). As a result, the number of replacements returned was
2761 double what it should be. I removed one of the increments, but Craig sent a
2762 later patch that removed the other one (the right fix) and added unit tests
2763 that check the return values (which was not done before).
2765 21. Several CMake things:
2767 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2768 the names libpcre and libpcreposix, not just pcre and pcreposix.
2770 (2) The above change means that pcretest and pcregrep are now correctly
2771 linked with the newly-built libraries, not previously installed ones.
2775 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2776 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2777 UTF-8 newline character). The key issue is that the pattern starts .*;
2778 this means that the match must be either at the beginning, or after a
2779 newline. The bug was in the code for advancing after a failed match and
2780 checking that the new position followed a newline. It was not taking
2781 account of UTF-8 characters correctly.
2783 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2784 character classes. PCRE was not treating the sequence [:...:] as a
2785 character class unless the ... were all letters. Perl, however, seems to
2786 allow any characters between [: and :], though of course it rejects as
2787 unknown any "names" that contain non-letters, because all the known class
2788 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2789 for example, whereas PCRE did not - it did not recognize a POSIX character
2790 class. This seemed a bit dangerous, so the code has been changed to be
2791 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2792 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2793 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2794 Perl does, and where it didn't before.
2796 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2797 Windows environments %n is disabled by default.
2800 Version 7.4 21-Sep-07
2801 ---------------------
2803 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2804 means that a class such as [\s] counted as "explicit reference to CR or
2805 LF". That isn't really right - the whole point of the change was to try to
2806 help when there was an actual mention of one of the two characters. So now
2807 the change happens only if \r or \n (or a literal CR or LF) character is
2808 encountered.
2810 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2811 of both had grown to the point where there were only 3 bits left.
2812 Fortunately, there was spare space in the data structure, and so I have
2813 moved the internal flags into a new 16-bit field to free up more option
2814 bits.
2816 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2817 but did not set the internal JCHANGED flag - either of these is enough to
2818 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2819 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2820 start sets both bits.
2822 4. Added options (at build time, compile time, exec time) to change \R from
2823 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2825 5. doc/pcresyntax.html was missing from the distribution.
2827 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2828 compatibility, even though it is no longer used.
2830 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2831 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2832 windows.h file is present (where different names are used). [This was
2833 reversed later after testing - see 16 below.]
2835 8. Changed all #include <config.h> to #include "config.h". There were also
2836 some further <pcre.h> cases that I changed to "pcre.h".
2838 9. When pcregrep was used with the --colour option, it missed the line ending
2839 sequence off the lines that it output.
2841 10. It was pointed out to me that arrays of string pointers cause lots of
2842 relocations when a shared library is dynamically loaded. A technique of
2843 using a single long string with a table of offsets can drastically reduce
2844 these. I have refactored PCRE in four places to do this. The result is
2845 dramatic:
2847 Originally: 290
2848 After changing UCP table: 187
2849 After changing error message table: 43
2850 After changing table of "verbs" 36
2851 After changing table of Posix names 22
2853 Thanks to the folks working on Gregex for glib for this insight.
2855 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2856 unicode-properties was also set.
2858 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2860 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2861 checked only for CRLF.
2863 14. Added casts to pcretest.c to avoid compiler warnings.
2865 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2867 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2868 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2869 entirely. This removes changes made in 7 above.
2871 17. The CMake files have been updated, and there is now more information about
2872 building with CMake in the NON-UNIX-USE document.
2875 Version 7.3 28-Aug-07
2876 ---------------------
2878 1. In the rejigging of the build system that eventually resulted in 7.1, the
2879 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2880 brackets there is not right, since it causes compilers to look for an
2881 installed pcre.h, not the version that is in the source that is being
2882 compiled (which of course may be different). I have changed it back to:
2884 #include "pcre.h"
2886 I have a vague recollection that the change was concerned with compiling in
2887 different directories, but in the new build system, that is taken care of
2888 by the VPATH setting the Makefile.
2890 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2891 when the subject happened to end in the byte 0x85 (e.g. if the last
2892 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2893 characters but of course it shouldn't be taken as a newline when it is part
2894 of another character. The bug was that, for an unlimited repeat of . in
2895 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2896 characters when looking for a newline.
2898 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2900 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2901 in debug output.
2903 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2904 long printing in the pcrecpp unittest when running under MinGW.
2906 6. ESC_K was left out of the EBCDIC table.
2908 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2909 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2910 limit also applies to "virtual nesting" when a pattern is recursive, and in
2911 this case 1000 isn't so big. I have been able to remove this limit at the
2912 expense of backing off one optimization in certain circumstances. Normally,
2913 when pcre_exec() would call its internal match() function recursively and
2914 immediately return the result unconditionally, it uses a "tail recursion"
2915 feature to save stack. However, when a subpattern that can match an empty
2916 string has an unlimited repetition quantifier, it no longer makes this
2917 optimization. That gives it a stack frame in which to save the data for
2918 checking that an empty string has been matched. Previously this was taken
2919 from the 1000-entry workspace that had been reserved. So now there is no
2920 explicit limit, but more stack is used.
2922 8. Applied Daniel's patches to solve problems with the import/export magic
2923 syntax that is required for Windows, and which was going wrong for the
2924 pcreposix and pcrecpp parts of the library. These were overlooked when this
2925 problem was solved for the main library.
2927 9. There were some crude static tests to avoid integer overflow when computing
2928 the size of patterns that contain repeated groups with explicit upper
2929 limits. As the maximum quantifier is 65535, the maximum group length was
2930 set at 30,000 so that the product of these two numbers did not overflow a
2931 32-bit integer. However, it turns out that people want to use groups that
2932 are longer than 30,000 bytes (though not repeat them that many times).
2933 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2934 made it possible to implement the integer overflow checks in a much more
2935 dynamic way, which I have now done. The artificial limitation on group
2936 length has been removed - we now have only the limit on the total length of
2937 the compiled pattern, which depends on the LINK_SIZE setting.
2939 10. Fixed a bug in the documentation for get/copy named substring when
2940 duplicate names are permitted. If none of the named substrings are set, the
2941 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2942 empty string.
2944 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2945 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2946 because the ] is interpreted as the first data character and the
2947 terminating ] is not found. PCRE has been made compatible with Perl in this
2948 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2949 cause memory overwriting.
2951 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2952 string has been matched (to stop an infinite loop). It was not recognizing
2953 a conditional subpattern that could match an empty string if that
2954 subpattern was within another subpattern. For example, it looped when
2955 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2956 condition was not nested. This bug has been fixed.
2958 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2959 past the start of the subject in the presence of bytes with the top bit
2960 set, for example "\x8aBCD".
2962 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2963 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2965 14. Optimized (?!) to (*FAIL).
2967 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2968 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2969 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2970 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2971 does: it's just the validity check that is more restrictive.
2973 16. Inserted checks for integer overflows during escape sequence (backslash)
2974 processing, and also fixed erroneous offset values for syntax errors during
2975 backslash processing.
2977 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2978 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2980 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2981 caused an overrun.
2983 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2984 something other than just ASCII characters) inside a group that had an
2985 unlimited repeat caused a loop at compile time (while checking to see
2986 whether the group could match an empty string).
2988 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2989 [\P{Any}] did so. (Error in the code for printing property names.)
2991 21. An orphan \E inside a character class could cause a crash.
2993 22. A repeated capturing bracket such as (A)? could cause a wild memory
2994 reference during compilation.
2996 23. There are several functions in pcre_compile() that scan along a compiled
2997 expression for various reasons (e.g. to see if it's fixed length for look
2998 behind). There were bugs in these functions when a repeated \p or \P was
2999 present in the pattern. These operators have additional parameters compared
3000 with \d, etc, and these were not being taken into account when moving along
3001 the compiled data. Specifically:
3003 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3004 length.
3006 (b) An item such as \pL+ within a repeated group could cause crashes or
3007 loops.
3009 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3010 "reference to non-existent subpattern" error.
3012 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3014 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3015 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3017 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3019 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3020 character were causing crashes (broken optimization).
3022 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3023 \p or \P) caused a compile-time loop.
3025 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3026 break. For example, the unstudied pattern [\r\n]A does not match the string
3027 "\r\nA" because change 7.0/46 below moves the current point on by two
3028 characters after failing to match at the start. However, the pattern \nA
3029 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3030 the same is true. There doesn't seem any very clean way out of this, but
3031 what I have chosen to do makes the common cases work: PCRE now takes note
3032 of whether there can be an explicit match for \r or \n anywhere in the
3033 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3034 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3035 pattern has explicit CR or LF references.
3037 29. Added (*CR) etc for changing newline setting at start of pattern.
3040 Version 7.2 19-Jun-07
3041 ---------------------
3043 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3044 which is apparently normally available under Windows.
3046 2. Re-jig the pcregrep tests with different newline settings in an attempt
3047 to make them independent of the local environment's newline setting.
3049 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3051 4. Some of the "internals" tests were previously cut out when the link size
3052 was not 2, because the output contained actual offsets. The recent new
3053 "Z" feature of pcretest means that these can be cut out, making the tests
3054 usable with all link sizes.
3056 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3057 stack recursion. This gives a massive performance boost under BSD, but just
3058 a small improvement under Linux. However, it saves one field in the frame
3059 in all cases.
3061 6. Added more features from the forthcoming Perl 5.10:
3063 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3064 recursion call. It refers to the nth most recently opened parentheses.
3066 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3067 to be opened parentheses.
3069 (c) Conditions that refer to capturing parentheses can be specified
3070 relatively, for example, (?(-2)... or (?(+3)...
3072 (d) \K resets the start of the current match so that everything before
3073 is not part of it.
3075 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3077 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3078 reference syntax.
3080 (g) (?| introduces a group in which the numbering of parentheses in each
3081 alternative starts with the same number.
3083 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3085 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3088 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3089 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3090 for detecting groups that can match an empty string.
3092 9. A pattern with a very large number of alternatives (more than several
3093 hundred) was running out of internal workspace during the pre-compile
3094 phase, where pcre_compile() figures out how much memory will be needed. A
3095 bit of new cunning has reduced the workspace needed for groups with
3096 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3097 workspace instead of running out of the 4096 that are available.
3099 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3101 11. Applied patch from Google to remove an optimization that didn't quite work.
3102 The report of the bug said:
3104 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3105 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3106 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3108 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3109 it matched the wrong number of bytes.
3112 Version 7.1 24-Apr-07
3113 ---------------------
3115 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3116 that is more "standard", making use of automake and other Autotools. There
3117 is some re-arrangement of the files and adjustment of comments consequent
3118 on this.
3120 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3121 for recursive directory scanning broke on some systems because the files
3122 are not scanned in any specific order and on different systems the order
3123 was different. A call to "sort" has been inserted into RunGrepTest for the
3124 approprate test as a short-term fix. In the longer term there may be an
3125 alternative.
3127 3. I had an email from Eric Raymond about problems translating some of PCRE's
3128 man pages to HTML (despite the fact that I distribute HTML pages, some
3129 people do their own conversions for various reasons). The problems
3130 concerned the use of low-level troff macros .br and .in. I have therefore
3131 removed all such uses from the man pages (some were redundant, some could
3132 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3133 HTML has been updated to handle .nf/.fi and to complain if it encounters
3134 .br or .in.
3136 4. Updated comments in configure.ac that get placed in config.h.in and also
3137 arranged for config.h to be included in the distribution, with the name
3138 config.h.generic, for the benefit of those who have to compile without
3139 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3141 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3142 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3143 makevp.bat for the current PCRE, using the additional files
3144 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3146 6. A Windows user reported a minor discrepancy with test 2, which turned out
3147 to be caused by a trailing space on an input line that had got lost in his
3148 copy. The trailing space was an accident, so I've just removed it.
3150 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3151 that is needed.
3153 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3154 as "const" (a) because they are and (b) because it helps the PHP
3155 maintainers who have recently made a script to detect big data structures
3156 in the php code that should be moved to the .rodata section. I remembered
3157 to update Builducptable as well, so it won't revert if ucptable.h is ever
3158 re-created.
3160 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3161 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3162 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3163 support is not required. This saves 1.5-2K of code, which is important in
3164 some applications.
3166 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3167 so as not to refer to the tables, even though these functions will never be
3168 called when UTF-8 support is disabled. Otherwise there are problems with a
3169 shared library.
3171 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3173 (a) It was defining its arguments as char * instead of void *.
3175 (b) It was assuming that all moves were upwards in memory; this was true
3176 a long time ago when I wrote it, but is no longer the case.
3178 The emulated memove() is provided for those environments that have neither
3179 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3180 is clearly not the case, as these two bugs were recently reported.
3182 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3183 and Detrail to create the HTML documentation, the .txt form of the man
3184 pages, and it removes trailing spaces from listed files. It also creates
3185 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3186 case, it wraps all the #defines with #ifndefs. This script should be run
3187 before "make dist".
3189 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3190 with Unicode property support.
3192 (a) For a maximizing quantifier, if the two different cases of the
3193 character were of different lengths in their UTF-8 codings (there are
3194 some cases like this - I found 11), and the matching function had to
3195 back up over a mixture of the two cases, it incorrectly assumed they
3196 were both the same length.
3198 (b) When PCRE was configured to use the heap rather than the stack for
3199 recursion during matching, it was not correctly preserving the data for
3200 the other case of a UTF-8 character when checking ahead for a match
3201 while processing a minimizing repeat. If the check also involved
3202 matching a wide character, but failed, corruption could cause an
3203 erroneous result when trying to check for a repeat of the original
3204 character.
3206 13. Some tidying changes to the testing mechanism:
3208 (a) The RunTest script now detects the internal link size and whether there
3209 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3210 values substituted by "configure". (The RunGrepTest script already did
3211 this for UTF-8.) The configure.ac script no longer substitutes the
3212 relevant variables.
3214 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3215 with length and offset values. This means that the output is different
3216 for different internal link sizes. Test 2 is skipped for link sizes
3217 other than 2 because of this, bypassing the problem. Unfortunately,
3218 there was also a test in test 3 (the locale tests) that used /B and
3219 failed for link sizes other than 2. Rather than cut the whole test out,
3220 I have added a new /Z option to pcretest that replaces the length and
3221 offset values with spaces. This is now used to make test 3 independent
3222 of link size. (Test 2 will be tidied up later.)
3224 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3225 segmentation fault instead of returning the appropriate error message.
3227 15. In multiline mode when the newline sequence was set to "any", the pattern
3228 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3229 This doesn't seem right; it now treats the CRLF combination as the line
3230 ending, and so does not match in that case. It's only a pattern such as ^$
3231 that would hit this one: something like ^ABC$ would have failed after \r
3232 and then tried again after \r\n.
3234 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3235 in an attempt to make files that differ only in their line terminators
3236 compare equal. This works on Linux.
3238 17. Under certain error circumstances pcregrep might try to free random memory
3239 as it exited. This is now fixed, thanks to valgrind.
3241 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3242 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3243 was because its rules for how to advance for /g after matching an empty
3244 string at the end of a line did not allow for this case. They now check for
3245 it specially.
3247 20. pcretest is supposed to handle patterns and data of any length, by
3248 extending its buffers when necessary. It was getting this wrong when the
3249 buffer for a data line had to be extended.
3251 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3252 CRLF as a newline sequence.
3254 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3255 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3256 I have nevertheless tidied it up.
3258 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3260 24. Added a man page for pcre-config.
3263 Version 7.0 19-Dec-06
3264 ---------------------
3266 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3267 moving to gcc 4.1.1.
3269 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3270 sys/time.h, which is documented as needed for this function. It doesn't
3271 seem to matter on Linux, but it showed up on some releases of OS X.
3273 3. It seems that there are systems where bytes whose values are greater than
3274 127 match isprint() in the "C" locale. The "C" locale should be the
3275 default when a C program starts up. In most systems, only ASCII printing
3276 characters match isprint(). This difference caused the output from pcretest
3277 to vary, making some of the tests fail. I have changed pcretest so that:
3279 (a) When it is outputting text in the compiled version of a pattern, bytes
3280 other than 32-126 are always shown as hex escapes.
3282 (b) When it is outputting text that is a matched part of a subject string,
3283 it does the same, unless a different locale has been set for the match
3284 (using the /L modifier). In this case, it uses isprint() to decide.
3286 4. Fixed a major bug that caused incorrect computation of the amount of memory
3287 required for a compiled pattern when options that changed within the
3288 pattern affected the logic of the preliminary scan that determines the
3289 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3290 that the computed length was too small. The symptoms of this bug were
3291 either the PCRE error "internal error: code overflow" from pcre_compile(),
3292 or a glibc crash with a message such as "pcretest: free(): invalid next
3293 size (fast)". Examples of patterns that provoked this bug (shown in
3294 pcretest format) are:
3296 /(?-x: )/x
3297 /(?x)(?-x: \s*#\s*)/
3298 /((?i)[\x{c0}])/8
3299 /(?i:[\x{c0}])/8
3301 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3302 is now done differently.
3304 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3305 wrapper classes; (b) implement a new function in the C++ scanner that is
3306 more efficient than the old way of doing things because it avoids levels of
3307 recursion in the regex matching; (c) add a paragraph to the documentation
3308 for the FullMatch() function.
3310 6. The escape sequence \n was being treated as whatever was defined as
3311 "newline". Not only was this contrary to the documentation, which states
3312 that \n is character 10 (hex 0A), but it also went horribly wrong when
3313 "newline" was defined as CRLF. This has been fixed.
3315 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3316 was being set to -1 for the "end of line" case (supposedly a value that no
3317 character can have). Though this value is never used (the check for end of
3318 line is "zero bytes in current character"), it caused compiler complaints.
3319 I've changed it to 0xffffffff.
3321 8. In pcre_version.c, the version string was being built by a sequence of
3322 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3323 string (as it is for production releases) called a macro with an empty
3324 argument. The C standard says the result of this is undefined. The gcc
3325 compiler treats it as an empty string (which was what was wanted) but it is
3326 reported that Visual C gives an error. The source has been hacked around to
3327 avoid this problem.
3329 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3330 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3331 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3332 of them did).
3334 10. Originally, pcretest opened its input and output without "b"; then I was
3335 told that "b" was needed in some environments, so it was added for release
3336 5.0 to both the input and output. (It makes no difference on Unix-like
3337 systems.) Later I was told that it is wrong for the input on Windows. I've
3338 now abstracted the modes into two macros, to make it easier to fiddle with
3339 them, and removed "b" from the input mode under Windows.
3341 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3343 12. Added -help and --help to pcretest as an official way of being reminded
3344 of the options.
3346 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3347 and pcrecpp.cc because they annoy compilers at high warning levels.
3349 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3351 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3352 scripts are not C programs :-) and which was not noticed because it works
3353 on Linux.
3355 16. pcretest is supposed to handle any length of pattern and data line (as one
3356 line or as a continued sequence of lines) by extending its input buffer if
3357 necessary. This feature was broken for very long pattern lines, leading to
3358 a string of junk being passed to pcre_compile() if the pattern was longer
3359 than about 50K.
3361 17. I have done a major re-factoring of the way pcre_compile() computes the
3362 amount of memory needed for a compiled pattern. Previously, there was code
3363 that made a preliminary scan of the pattern in order to do this. That was
3364 OK when PCRE was new, but as the facilities have expanded, it has become
3365 harder and harder to keep it in step with the real compile phase, and there
3366 have been a number of bugs (see for example, 4 above). I have now found a
3367 cunning way of running the real compile function in a "fake" mode that
3368 enables it to compute how much memory it would need, while actually only
3369 ever using a few hundred bytes of working memory and without too many
3370 tests of the mode. This should make future maintenance and development
3371 easier. A side effect of this work is that the limit of 200 on the nesting
3372 depth of parentheses has been removed (though this was never a serious
3373 limitation, I suspect). However, there is a downside: pcre_compile() now
3374 runs more slowly than before (30% or more, depending on the pattern). I
3375 hope this isn't a big issue. There is no effect on runtime performance.
3377 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3378 newline (only possible for the last line of a file) and it was a
3379 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3381 19. Added additional timing features to pcretest. (1) The -tm option now times
3382 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3383 separate command line item, by a number that specifies the number of
3384 repeats to use when timing. The default is 50000; this gives better
3385 precision, but takes uncomfortably long for very large patterns.
3387 20. Extended pcre_study() to be more clever in cases where a branch of a
3388 subpattern has no definite first character. For example, (a*|b*)[cd] would
3389 previously give no result from pcre_study(). Now it recognizes that the
3390 first character must be a, b, c, or d.
3392 21. There was an incorrect error "recursive call could loop indefinitely" if
3393 a subpattern (or the entire pattern) that was being tested for matching an
3394 empty string contained only one non-empty item after a nested subpattern.
3395 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3396 incorrectly, because the \d was being skipped in the check.
3398 22. The pcretest program now has a new pattern option /B and a command line
3399 option -b, which is equivalent to adding /B to every pattern. This causes
3400 it to show the compiled bytecode, without the additional information that
3401 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3402 is the same as /B/I).
3404 23. A new optimization is now able automatically to treat some sequences such
3405 as a*b as a*+b. More specifically, if something simple (such as a character
3406 or a simple class like \d) has an unlimited quantifier, and is followed by
3407 something that cannot possibly match the quantified thing, the quantifier
3408 is automatically "possessified".
3410 24. A recursive reference to a subpattern whose number was greater than 39
3411 went wrong under certain circumstances in UTF-8 mode. This bug could also
3412 have affected the operation of pcre_study().
3414 25. Realized that a little bit of performance could be had by replacing
3415 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3417 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3419 27. Possessive quantifiers such as a++ were previously implemented by turning
3420 them into atomic groups such as ($>a+). Now they have their own opcodes,
3421 which improves performance. This includes the automatically created ones
3422 from 23 above.
3424 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3425 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3426 the first matched character to be a colon. This applied both to named and
3427 numbered groups.
3429 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3431 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3432 building PCRE on Windows easier, so I have included it in the distribution.
3434 31. There is now a check in pcretest against a ridiculously large number being
3435 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3436 loop, the loop is abandoned.
3438 32. Forward references to subpatterns in conditions such as (?(2)...) where
3439 subpattern 2 is defined later cause pcre_compile() to search forwards in
3440 the pattern for the relevant set of parentheses. This search went wrong
3441 when there were unescaped parentheses in a character class, parentheses
3442 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3444 33. "Subroutine" calls and backreferences were previously restricted to
3445 referencing subpatterns earlier in the regex. This restriction has now
3446 been removed.
3448 34. Added a number of extra features that are going to be in Perl 5.10. On the
3449 whole, these are just syntactic alternatives for features that PCRE had
3450 previously implemented using the Python syntax or my own invention. The
3451 other formats are all retained for compatibility.
3453 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3454 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3455 also .NET compatible.
3457 (b) A recursion or subroutine call to a named group can now be defined as
3458 (?&name) as well as (?P>name).
3460 (c) A backreference to a named group can now be defined as \k<name> or
3461 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3462 5.10, are also .NET compatible.
3464 (d) A conditional reference to a named group can now use the syntax
3465 (?(<name>) or (?('name') as well as (?(name).
3467 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3468 groups (named and numbered) that are never evaluated inline, but can be
3469 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3470 is always false. There may be only one alternative in such a group.
3472 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3473 as the simple (?(R). The condition is true only if the most recent
3474 recursion is that of the given number or name. It does not search out
3475 through the entire recursion stack.
3477 (g) The escape \gN or \g{N} has been added, where N is a positive or
3478 negative number, specifying an absolute or relative reference.
3480 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3481 some "unreachable code" warnings.
3483 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3484 things, this adds five new scripts.
3486 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3487 There were also incompatibilities regarding the handling of \Q..\E inside
3488 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3489 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3491 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3492 matches an empty string, and forcibly breaks the loop. There were bugs in
3493 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3494 against aaaa the result was just "a" rather than "aaaa", for example. Two
3495 separate and independent bugs (that affected different cases) have been
3496 fixed.
3498 39. Refactored the code to abolish the use of different opcodes for small
3499 capturing bracket numbers. This is a tidy that I avoided doing when I
3500 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3501 The new approach is not only tidier, it makes it possible to reduce the
3502 memory needed to fix the previous bug (38).
3504 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3505 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3506 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3507 mode.
3509 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3510 report.
3512 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3513 copy construction and assignment in the C++ wrapper.
3515 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3516 couple of bugs that could have given wrong results in the "--newline=crlf"
3517 case.
3519 44. Added a number of casts and did some reorganization of signed/unsigned int
3520 variables following suggestions from Dair Grant. Also renamed the variable
3521 "this" as "item" because it is a C++ keyword.
3523 45. Arranged for dftables to add
3525 #include "pcre_internal.h"
3527 to pcre_chartables.c because without it, gcc 4.x may remove the array
3528 definition from the final binary if PCRE is built into a static library and
3529 dead code stripping is activated.
3531 46. For an unanchored pattern, if a match attempt fails at the start of a
3532 newline sequence, and the newline setting is CRLF or ANY, and the next two
3533 characters are CRLF, advance by two characters instead of one.
3536 Version 6.7 04-Jul-06
3537 ---------------------
3539 1. In order to handle tests when input lines are enormously long, pcretest has
3540 been re-factored so that it automatically extends its buffers when
3541 necessary. The code is crude, but this _is_ just a test program. The
3542 default size has been increased from 32K to 50K.
3544 2. The code in pcre_study() was using the value of the re argument before
3545 testing it for NULL. (Of course, in any sensible call of the function, it
3546 won't be NULL.)
3548 3. The memmove() emulation function in pcre_internal.h, which is used on
3549 systems that lack both memmove() and bcopy() - that is, hardly ever -
3550 was missing a "static" storage class specifier.
3552 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3553 containing an extended class (one that cannot be represented by a bitmap
3554 because it contains high-valued characters or Unicode property items, e.g.
3555 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3556 pattern, but PCRE should not loop if you do not (it no longer does).
3557 [Detail: two cases were found: (a) a repeated subpattern containing an
3558 extended class; (b) a recursive reference to a subpattern that followed a
3559 previous extended class. It wasn't skipping over the extended class
3560 correctly when UTF-8 mode was not set.]
3562 5. A negated single-character class was not being recognized as fixed-length
3563 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3564 compile error "lookbehind assertion is not fixed length".
3566 6. The RunPerlTest auxiliary script was showing an unexpected difference
3567 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3568 write a Perl script that can interpret lines of an input file either as
3569 byte characters or as UTF-8, which is what "perltest" was being required to
3570 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3571 can't do is switch easily at run time between having the "use utf8;" pragma
3572 or not. In the end, I fudged it by using the RunPerlTest script to insert
3573 "use utf8;" explicitly for the UTF-8 tests.
3575 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3576 the end of the subject string, contrary to the documentation and to what
3577 Perl does. This was true of both matching functions. Now it matches only at
3578 the start of the subject and immediately after *internal* newlines.
3580 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3581 a pointer to an int instead of a pointer to an unsigned long int. This
3582 caused problems on 64-bit systems.
3584 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3585 instance of the 'standard' template library not being so standard".
3587 10. There was no check on the number of named subpatterns nor the maximum
3588 length of a subpattern name. The product of these values is used to compute
3589 the size of the memory block for a compiled pattern. By supplying a very
3590 long subpattern name and a large number of named subpatterns, the size
3591 computation could be caused to overflow. This is now prevented by limiting
3592 the length of names to 32 characters, and the number of named subpatterns
3593 to 10,000.
3595 11. Subpatterns that are repeated with specific counts have to be replicated in
3596 the compiled pattern. The size of memory for this was computed from the
3597 length of the subpattern and the repeat count. The latter is limited to
3598 65535, but there was no limit on the former, meaning that integer overflow
3599 could in principle occur. The compiled length of a repeated subpattern is
3600 now limited to 30,000 bytes in order to prevent this.
3602 12. Added the optional facility to have named substrings with the same name.
3604 13. Added the ability to use a named substring as a condition, using the
3605 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3606 are numbers (not recommended). Forward references are permitted.
3608 14. Added forward references in named backreferences (if you see what I mean).
3610 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3611 pattern could run off the end of the subject. For example, the pattern
3612 "(?s)(.{1,5})"8 did this with the subject "ab".
3614 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3615 PCRE_CASELESS was set when matching characters that were quantified with ?
3616 or *.
3618 17. A character class other than a single negated character that had a minimum
3619 but no maximum quantifier - for example [ab]{6,} - was not handled
3620 correctly by pce_dfa_exec(). It would match only one character.
3622 18. A valid (though odd) pattern that looked like a POSIX character
3623 class but used an invalid character after [ (for example [[,abc,]]) caused
3624 pcre_compile() to give the error "Failed: internal error: code overflow" or
3625 in some cases to crash with a glibc free() error. This could even happen if
3626 the pattern terminated after [[ but there just happened to be a sequence of
3627 letters, a binary zero, and a closing ] in the memory that followed.
3629 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3630 over the years. Originally (before any Unicode support), just the bottom 8
3631 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3632 output from "man perlunicode" includes this:
3634 The regular expression compiler produces polymorphic opcodes. That
3635 is, the pattern adapts to the data and automatically switches to
3636 the Unicode character scheme when presented with Unicode data--or
3637 instead uses a traditional byte scheme when presented with byte
3638 data.
3640 Sadly, a wide octal escape does not cause a switch, and in a string with
3641 no other multibyte characters, these octal escapes are treated as before.
3642 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3643 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3644 Unicode string.
3646 I have not perpetrated such confusion in PCRE. Up till now, it took just
3647 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3648 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3649 translate to the appropriate multibyte character.
3651 29. Applied some refactoring to reduce the number of warnings from Microsoft
3652 and Borland compilers. This has included removing the fudge introduced
3653 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3654 a warning about an unused variable.
3656 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3657 characters since release 4.0, because Perl (from release 5.004) does not.
3658 [Or at least, is documented not to: some releases seem to be in conflict
3659 with the documentation.] However, when a pattern was studied with
3660 pcre_study() and all its branches started with \s, PCRE still included VT
3661 as a possible starting character. Of course, this did no harm; it just
3662 caused an unnecessary match attempt.
3664 22. Removed a now-redundant internal flag bit that recorded the fact that case
3665 dependency changed within the pattern. This was once needed for "required
3666 byte" processing, but is no longer used. This recovers a now-scarce options
3667 bit. Also moved the least significant internal flag bit to the most-
3668 significant bit of the word, which was not previously used (hangover from
3669 the days when it was an int rather than a uint) to free up another bit for
3670 the future.
3672 23. Added support for CRLF line endings as well as CR and LF. As well as the
3673 default being selectable at build time, it can now be changed at runtime
3674 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3675 specify that it is scanning data with non-default line endings.
3677 24. Changed the definition of CXXLINK to make it agree with the definition of
3678 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3680 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3681 recursions. This makes a big different to stack usage for some patterns.
3683 26. If a subpattern containing a named recursion or subroutine reference such
3684 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3685 the space required for the compiled pattern went wrong and gave too small a
3686 value. Depending on the environment, this could lead to "Failed: internal
3687 error: code overflow at offset 49" or "glibc detected double free or
3688 corruption" errors.
3690 27. Applied patches from Google (a) to support the new newline modes and (b) to
3691 advance over multibyte UTF-8 characters in GlobalReplace.
3693 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3694 difference for some implementation of PCRE in some Windows version.
3696 29. Added some extra testing facilities to pcretest:
3698 \q<number> in a data line sets the "match limit" value
3699 \Q<number> in a data line sets the "match recursion limt" value
3700 -S <number> sets the stack size, where <number> is in megabytes
3702 The -S option isn't available for Windows.
3705 Version 6.6 06-Feb-06
3706 ---------------------
3708 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3709 in pcreposix.h. I have copied the definition from pcre.h.
3711 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3712 because pcre.h is no longer a built file.
3714 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3715 not normally included in the compiled code.
3718 Version 6.5 01-Feb-06
3719 ---------------------
3721 1. When using the partial match feature with pcre_dfa_exec(), it was not
3722 anchoring the second and subsequent partial matches at the new starting
3723 point. This could lead to incorrect results. For example, with the pattern
3724 /1234/, partially matching against "123" and then "a4" gave a match.
3726 2. Changes to pcregrep:
3728 (a) All non-match returns from pcre_exec() were being treated as failures
3729 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3730 error message is output. Some extra information is given for the
3732 probably the only errors that are likely to be caused by users (by
3733 specifying a regex that has nested indefinite repeats, for instance).
3734 If there are more than 20 of these errors, pcregrep is abandoned.
3736 (b) A binary zero was treated as data while matching, but terminated the
3737 output line if it was written out. This has been fixed: binary zeroes
3738 are now no different to any other data bytes.
3740 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3741 used to set a locale for matching. The --locale=xxxx long option has
3742 been added (no short equivalent) to specify a locale explicitly on the
3743 pcregrep command, overriding the environment variables.
3745 (d) When -B was used with -n, some line numbers in the output were one less
3746 than they should have been.
3748 (e) Added the -o (--only-matching) option.
3750 (f) If -A or -C was used with -c (count only), some lines of context were
3751 accidentally printed for the final match.
3753 (g) Added the -H (--with-filename) option.
3755 (h) The combination of options -rh failed to suppress file names for files
3756 that were found from directory arguments.
3758 (i) Added the -D (--devices) and -d (--directories) options.
3760 (j) Added the -F (--fixed-strings) option.
3762 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3764 (l) Added the --colo(u)r option.
3766 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3767 is not present by default.
3769 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3770 items such as (?R) or (?1), when the recursion could match a number of
3771 alternatives. If it matched one of the alternatives, but subsequently,
3772 outside the recursion, there was a failure, the code tried to back up into
3773 the recursion. However, because of the way PCRE is implemented, this is not
3774 possible, and the result was an incorrect result from the match.
3776 In order to prevent this happening, the specification of recursion has
3777 been changed so that all such subpatterns are automatically treated as
3778 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3780 4. I had overlooked the fact that, in some locales, there are characters for
3781 which isalpha() is true but neither isupper() nor islower() are true. In
3782 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3783 and ordfeminine) are like this. This affected the treatment of \w and \W
3784 when they appeared in character classes, but not when they appeared outside
3785 a character class. The bit map for "word" characters is now created
3786 separately from the results of isalnum() instead of just taking it from the
3787 upper, lower, and digit maps. (Plus the underscore character, of course.)
3789 5. The above bug also affected the handling of POSIX character classes such as
3790 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3791 permanent tables. Instead, the bit maps for such a class were previously
3792 created as the appropriate unions of the upper, lower, and digit bitmaps.
3793 Now they are created by subtraction from the [[:word:]] class, which has
3794 its own bitmap.
3796 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3797 It is created by subtracting the vertical space characters (\x09, \x0a,
3798 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3799 subtraction was done in the overall bitmap for a character class, meaning
3800 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3801 be recognized. This bug has been fixed.
3803 7. Patches from the folks at Google:
3805 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3806 real life, but is still worth protecting against".
3808 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3809 regular expressions".
3811 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3812 have it.
3814 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3815 "configure" and the latter not, in order to fix a problem somebody had
3816 with compiling the Arg class on HP-UX.
3818 (e) Improve the error-handling of the C++ wrapper a little bit.
3820 (f) New tests for checking recursion limiting.
3822 8. The pcre_memmove() function, which is used only if the environment does not
3823 have a standard memmove() function (and is therefore rarely compiled),
3824 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3825 returning a result (though PCRE never actually uses the result).
3827 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3828 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3829 returned instead of calling malloc() with an overflowing number that would
3830 most likely cause subsequent chaos.
3832 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3834 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3835 with this option is matched, the nmatch and pmatch options of regexec() are
3836 ignored.
3838 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3839 provided in case anyone wants to the the POSIX interface with UTF-8
3840 strings.
3842 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3843 C++ linking (needed for some HP-UX environments).
3845 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3846 (unused parameter) and in the pcre_printint() function (omitted "default"
3847 switch label when the default is to do nothing).
3849 15. Added some code to make it possible, when PCRE is compiled as a C++
3850 library, to replace subject pointers for pcre_exec() with a smart pointer
3851 class, thus making it possible to process discontinuous strings.
3853 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3854 much the same function. They were added by different people who were trying
3855 to make PCRE easy to compile on non-Unix systems. It has been suggested
3856 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3857 for compiling on Windows systems. I have therefore replaced it with
3858 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3859 defaults to "extern" for C or "extern C" for C++, which works fine on
3860 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3861 SCOPE with something explicit in config.h. In addition:
3863 (a) pcreposix.h still had just "extern" instead of either of these macros;
3864 I have replaced it with PCRE_DATA_SCOPE.
3866 (b) Functions such as _pcre_xclass(), which are internal to the library,
3867 but external in the C sense, all had PCRE_EXPORT in their definitions.
3868 This is apparently wrong for the Windows case, so I have removed it.
3869 (It makes no difference on Unix-like systems.)
3871 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3872 of recursive calls to match(). This is different to MATCH_LIMIT because
3873 that limits the total number of calls to match(), not all of which increase
3874 the depth of recursion. Limiting the recursion depth limits the amount of
3875 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3876 when PCRE is compiled, and changed at run time. A patch from Google adds
3877 this functionality to the C++ interface.
3879 18. Changes to the handling of Unicode character properties:
3881 (a) Updated the table to Unicode 4.1.0.
3883 (b) Recognize characters that are not in the table as "Cn" (undefined).
3885 (c) I revised the way the table is implemented to a much improved format
3886 which includes recognition of ranges. It now supports the ranges that
3887 are defined in UnicodeData.txt, and it also amalgamates other
3888 characters into ranges. This has reduced the number of entries in the
3889 table from around 16,000 to around 3,000, thus reducing its size
3890 considerably. I realized I did not need to use a tree structure after
3891 all - a binary chop search is just as efficient. Having reduced the
3892 number of entries, I extended their size from 6 bytes to 8 bytes to
3893 allow for more data.
3895 (d) Added support for Unicode script names via properties such as \p{Han}.
3897 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3898 matching that character.
3900 20. When matching a repeated Unicode property with a minimum greater than zero,
3901 (for example \pL{2,}), PCRE could look past the end of the subject if it
3902 reached it while seeking the minimum number of characters. This could
3903 happen only if some of the characters were more than one byte long, because
3904 there is a check for at least the minimum number of bytes.
3906 21. Refactored the implementation of \p and \P so as to be more general, to
3907 allow for more different types of property in future. This has changed the
3908 compiled form incompatibly. Anybody with saved compiled patterns that use
3909 \p or \P will have to recompile them.
3911 22. Added "Any" and "L&" to the supported property types.
3913 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3914 but give a compile time error if the value is greater than 0xff.
3916 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3917 accidentally not being installed or uninstalled.
3919 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3920 made were to insert the current release number. This seemed silly, because
3921 it made things harder for people building PCRE on systems that don't run
3922 "configure". I have turned pcre.h into a distributed file, no longer built
3923 by "configure", with the version identification directly included. There is
3924 no longer a pcre.h.in file.
3926 However, this change necessitated a change to the pcre-config script as
3927 well. It is built from pcre-config.in, and one of the substitutions was the
3928 release number. I have updated configure.ac so that ./configure now finds
3929 the release number by grepping pcre.h.
3931 26. Added the ability to run the tests under valgrind.
3934 Version 6.4 05-Sep-05
3935 ---------------------
3937 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3938 "--" to be printed when multiple files were scanned, even when none of the
3939 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3940 consider it to be a bug, and have restored the previous behaviour.
3942 2. A couple of code tidies to get rid of compiler warnings.
3944 3. The pcretest program used to cheat by referring to symbols in the library
3945 whose names begin with _pcre_. These are internal symbols that are not
3946 really supposed to be visible externally, and in some environments it is
3947 possible to suppress them. The cheating is now confined to including
3948 certain files from the library's source, which is a bit cleaner.
3950 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3951 file's purpose clearer.
3953 5. Reorganized pcre_ucp_findchar().
3956 Version 6.3 15-Aug-05
3957 ---------------------
3959 1. The file libpcre.pc.in did not have general read permission in the tarball.
3961 2. There were some problems when building without C++ support:
3963 (a) If C++ support was not built, "make install" and "make test" still
3964 tried to test it.
3966 (b) There were problems when the value of CXX was explicitly set. Some
3967 changes have been made to try to fix these, and ...
3969 (c) --disable-cpp can now be used to explicitly disable C++ support.
3971 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3972 backslash in a target when C++ was disabled. This confuses some
3973 versions of "make", apparently. Using an intermediate variable solves
3974 this. (Same for CPP_LOBJ.)
3976 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3977 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3978 necessary on certain architectures.
3980 4. Added a setting of -export-symbols-regex to the link command to remove
3981 those symbols that are exported in the C sense, but actually are local
3982 within the library, and not documented. Their names all begin with
3983 "_pcre_". This is not a perfect job, because (a) we have to except some
3984 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3985 available (and never for static libraries). I have made a note to try to
3986 find a way round (a) in the future.
3989 Version 6.2 01-Aug-05
3990 ---------------------
3992 1. There was no test for integer overflow of quantifier values. A construction
3993 such as {1111111111111111} would give undefined results. What is worse, if
3994 a minimum quantifier for a parenthesized subpattern overflowed and became
3995 negative, the calculation of the memory size went wrong. This could have
3996 led to memory overwriting.
3998 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4000 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4001 operating environments where this matters.
4003 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4004 PCRE options from within the C++ wrapper.
4006 5. Named capturing subpatterns were not being correctly counted when a pattern
4007 was compiled. This caused two problems: (a) If there were more than 100
4008 such subpatterns, the calculation of the memory needed for the whole
4009 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4010 back references of the form \12, where the number was greater than 9, were
4011 not recognized as back references, even though there were sufficient
4012 previous subpatterns.
4014 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4015 versions of gcc, e.g. 2.95.4.
4018 Version 6.1 21-Jun-05
4019 ---------------------
4021 1. There was one reference to the variable "posix" in pcretest.c that was not
4022 surrounded by "#if !defined NOPOSIX".
4024 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4025 the cross-check on the old pcre_info() function, for the benefit of the
4026 cut-down version of PCRE that is currently imported into Exim.
4028 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4029 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4030 patterns that start (?i) but I don't think that matters. The use of (?i) is
4031 just an example; this all applies to the other options as well.
4033 4. Since libtool seems to echo the compile commands it is issuing, the output
4034 from "make" can be reduced a bit by putting "@" in front of each libtool
4035 compile command.
4037 5. Patch from the folks at Google for configure.in to be a bit more thorough
4038 in checking for a suitable C++ installation before trying to compile the
4039 C++ stuff. This should fix a reported problem when a compiler was present,
4040 but no suitable headers.
4042 6. The man pages all had just "PCRE" as their title. I have changed them to
4043 be the relevant file name. I have also arranged that these names are
4044 retained in the file doc/pcre.txt, which is a concatenation in text format
4045 of all the man pages except the little individual ones for each function.
4047 7. The NON-UNIX-USE file had not been updated for the different set of source
4048 files that come with release 6. I also added a few comments about the C++
4049 wrapper.
4052 Version 6.0 07-Jun-05
4053 ---------------------
4055 1. Some minor internal re-organization to help with my DFA experiments.
4057 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4058 didn't matter for the library itself when fully configured, but did matter
4059 when compiling without UCP support, or within Exim, where the ucp files are
4060 not imported.
4062 3. Refactoring of the library code to split up the various functions into
4063 different source modules. The addition of the new DFA matching code (see
4064 below) to a single monolithic source would have made it really too
4065 unwieldy, quite apart from causing all the code to be include in a
4066 statically linked application, when only some functions are used. This is
4067 relevant even without the DFA addition now that patterns can be compiled in
4068 one application and matched in another.
4070 The downside of splitting up is that there have to be some external
4071 functions and data tables that are used internally in different modules of
4072 the library but which are not part of the API. These have all had their
4073 names changed to start with "_pcre_" so that they are unlikely to clash
4074 with other external names.
4076 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4077 a different (DFA) algorithm. Although it is slower than the original
4078 function, it does have some advantages for certain types of matching
4079 problem.
4081 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4082 including restarting after a partial match.
4084 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4085 defined when compiling for Windows was sent to me. I have put it into the
4086 code, though I have no means of testing or verifying it.
4088 7. Added the pcre_refcount() auxiliary function.
4090 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4091 match before or at the first newline in the subject string. In pcretest,
4092 the /f option on a pattern can be used to set this.
4094 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4095 would behave wrongly. This has been present in PCRE since release 4.0.
4097 10. A number of changes to the pcregrep command:
4099 (a) Refactored how -x works; insert ^(...)$ instead of setting
4100 PCRE_ANCHORED and checking the length, in preparation for adding
4101 something similar for -w.
4103 (b) Added the -w (match as a word) option.
4105 (c) Refactored the way lines are read and buffered so as to have more
4106 than one at a time available.
4108 (d) Implemented a pcregrep test script.
4110 (e) Added the -M (multiline match) option. This allows patterns to match
4111 over several lines of the subject. The buffering ensures that at least
4112 8K, or the rest of the document (whichever is the shorter) is available
4113 for matching (and similarly the previous 8K for lookbehind assertions).
4115 (f) Changed the --help output so that it now says
4117 -w, --word-regex(p)
4119 instead of two lines, one with "regex" and the other with "regexp"
4120 because that confused at least one person since the short forms are the
4121 same. (This required a bit of code, as the output is generated
4122 automatically from a table. It wasn't just a text change.)
4124 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4125 option but starts with a hyphen. Could be a pattern or a path name
4126 starting with a hyphen, for instance.
4128 (h) "-" can be given as a file name to represent stdin.
4130 (i) When file names are being printed, "(standard input)" is used for
4131 the standard input, for compatibility with GNU grep. Previously
4132 "<stdin>" was used.
4134 (j) The option --label=xxx can be used to supply a name to be used for
4135 stdin when file names are being printed. There is no short form.
4137 (k) Re-factored the options decoding logic because we are going to add
4138 two more options that take data. Such options can now be given in four
4139 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4141 (l) Added the -A, -B, and -C options for requesting that lines of context
4142 around matches be printed.
4144 (m) Added the -L option to print the names of files that do not contain
4145 any matching lines, that is, the complement of -l.
4147 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4148 continue to scan other files.
4150 (o) The -s option was incorrectly implemented. For compatibility with other
4151 greps, it now suppresses the error message for a non-existent or non-
4152 accessible file (but not the return code). There is a new option called
4153 -q that suppresses the output of matching lines, which was what -s was
4154 previously doing.
4156 (p) Added --include and --exclude options to specify files for inclusion
4157 and exclusion when recursing.
4159 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4160 Hopefully, it now does.
4162 12. Missing cast in pcre_study().
4164 13. Added an "uninstall" target to the makefile.
4166 14. Replaced "extern" in the function prototypes in Makefile.in with
4167 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4168 world, but is set differently for Windows.
4170 15. Added a second compiling function called pcre_compile2(). The only
4171 difference is that it has an extra argument, which is a pointer to an
4172 integer error code. When there is a compile-time failure, this is set
4173 non-zero, in addition to the error test pointer being set to point to an
4174 error message. The new argument may be NULL if no error number is required
4175 (but then you may as well call pcre_compile(), which is now just a
4176 wrapper). This facility is provided because some applications need a
4177 numeric error indication, but it has also enabled me to tidy up the way
4178 compile-time errors are handled in the POSIX wrapper.
4180 16. Added VPATH=.libs to the makefile; this should help when building with one
4181 prefix path and installing with another. (Or so I'm told by someone who
4182 knows more about this stuff than I do.)
4184 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4185 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4186 match everything, including newlines. This is not POSIX-compatible, but
4187 somebody wanted the feature. From pcretest it can be activated by using
4188 both the P and the s flags.
4190 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4192 19. libpcre.pc was being incorrectly installed as executable.
4194 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4195 it now also looks for '\r' so that it will work unmodified on Windows.
4197 21. Added Google's contributed C++ wrapper to the distribution.
4199 22. Added some untidy missing memory free() calls in pcretest, to keep
4200 Electric Fence happy when testing.
4204 Version 5.0 13-Sep-04
4205 ---------------------
4207 1. Internal change: literal characters are no longer packed up into items
4208 containing multiple characters in a single byte-string. Each character
4209 is now matched using a separate opcode. However, there may be more than one
4210 byte in the character in UTF-8 mode.
4212 2. The pcre_callout_block structure has two new fields: pattern_position and
4213 next_item_length. These contain the offset in the pattern to the next match
4214 item, and its length, respectively.
4216 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4217 insertion of callouts before each pattern item. Added the /C option to
4218 pcretest to make use of this.
4220 4. On the advice of a Windows user, the lines
4222 #if defined(_WIN32) || defined(WIN32)
4223 _setmode( _fileno( stdout ), 0x8000 );
4224 #endif /* defined(_WIN32) || defined(WIN32) */
4226 have been added to the source of pcretest. This apparently does useful
4227 magic in relation to line terminators.
4229 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4230 for the benefit of those environments where the "b" makes a difference.
4232 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4233 to know about it. I have put a hack into configure.in that adds in code
4234 to set GCC=yes if CC=icc. This seems to end up at a point in the
4235 generated configure script that is early enough to affect the setting of
4236 compiler options, which is what is needed, but I have no means of testing
4237 whether it really works. (The user who reported this had patched the
4238 generated configure script, which of course I cannot do.)
4240 LATER: After change 22 below (new libtool files), the configure script
4241 seems to know about icc (and also ecc). Therefore, I have commented out
4242 this hack in configure.in.
4244 7. Added support for pkg-config (2 patches were sent in).
4246 8. Negated POSIX character classes that used a combination of internal tables
4247 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4248 [[:^ascii]]. Typically, they would match almost any characters. The other
4249 POSIX classes were not broken in this way.
4251 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4252 to find the match, as PCRE was deluded into thinking that the match had to
4253 start at the start point or following a newline. The same bug applied to
4254 patterns with negative forward assertions or any backward assertions
4255 preceding ".*" at the start, unless the pattern required a fixed first
4256 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4258 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4259 starting at the last subject character, bytes beyond the end of the subject
4260 string were read.
4262 11. Renamed the variable "class" as "classbits" to make life easier for C++
4263 users. (Previously there was a macro definition, but it apparently wasn't
4264 enough.)
4266 12. Added the new field "tables" to the extra data so that tables can be passed
4267 in at exec time, or the internal tables can be re-selected. This allows
4268 a compiled regex to be saved and re-used at a later time by a different
4269 program that might have everything at different addresses.
4271 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4272 -R library as well as a -L library.
4274 14. The debugging options of pcretest (-d on the command line or D on a
4275 pattern) showed incorrect output for anything following an extended class
4276 that contained multibyte characters and which was followed by a quantifier.
4278 15. Added optional support for general category Unicode character properties
4279 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4280 support. It adds about 90K to the size of the library. The meanings of the
4281 inbuilt class escapes such as \d and \s have NOT been changed.
4283 16. Updated pcredemo.c to include calls to free() to release the memory for the
4284 compiled pattern.
4286 17. The generated file chartables.c was being created in the source directory
4287 instead of in the building directory. This caused the build to fail if the
4288 source directory was different from the building directory, and was
4289 read-only.
4291 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4292 file. No doubt somebody will tell me if they don't make sense... Also added
4293 Dan Mooney's comments about building on OpenVMS.
4295 19. Added support for partial matching via the PCRE_PARTIAL option for
4296 pcre_exec() and the \P data escape in pcretest.
4298 20. Extended pcretest with 3 new pattern features:
4300 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4301 write the compiled pattern to the file whose name is "rest-of-line".
4302 This is a straight binary dump of the data, with the saved pointer to
4303 the character tables forced to be NULL. The study data, if any, is
4304 written too. After writing, pcretest reads a new pattern.
4306 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4307 compiled pattern from the given file. There must not be any
4308 occurrences of "<" in the file name (pretty unlikely); if there are,
4309 pcretest will instead treat the initial "<" as a pattern delimiter.
4310 After reading in the pattern, pcretest goes on to read data lines as
4311 usual.
4313 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4314 and 16-bit fields in a compiled pattern, to simulate a pattern that
4315 was compiled on a host of opposite endianness.
4317 21. The pcre-exec() function can now cope with patterns that were compiled on
4318 hosts of opposite endianness, with this restriction:
4320 As for any compiled expression that is saved and used later, the tables
4321 pointer field cannot be preserved; the extra_data field in the arguments
4322 to pcre_exec() should be used to pass in a tables address if a value
4323 other than the default internal tables were used at compile time.
4325 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4326 now diagnosed as an error. Previously, most of the time, a negative number
4327 would have been treated as zero, but if in addition "ovector" was passed as
4328 NULL, a crash could occur.
4330 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4331 new versions from the libtool 1.5 distribution (the last one is a copy of
4332 a file called libtool.m4). This seems to have fixed the need to patch
4333 "configure" to support Darwin 1.3 (which I used to do). However, I still
4334 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4335 workstation).
4337 24. Changed the PCRE licence to be the more standard "BSD" licence.
4340 Version 4.5 01-Dec-03
4341 ---------------------
4343 1. There has been some re-arrangement of the code for the match() function so
4344 that it can be compiled in a version that does not call itself recursively.
4345 Instead, it keeps those local variables that need separate instances for
4346 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4347 needs to "recurse". Keeping track of where control must go is done by means
4348 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4349 hide most of the details from the main code, and operates only if
4350 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4351 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4352 operating.
4354 To make it easier for callers to provide specially tailored get/free
4355 functions for this usage, two new functions, pcre_stack_malloc, and
4356 pcre_stack_free, are used. They are always called in strict stacking order,
4357 and the size of block requested is always the same.
4359 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4360 PCRE has been compiled to use the stack or the heap for recursion. The
4361 -C option of pcretest uses this to show which version is compiled.
4363 A new data escape \S, is added to pcretest; it causes the amounts of store
4364 obtained and freed by both kinds of malloc/free at match time to be added
4365 to the output.
4367 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4368 what's available on my current Linux desktop machine.
4370 3. When matching a UTF-8 string, the test for a valid string at the start has
4371 been extended. If start_offset is not zero, PCRE now checks that it points
4372 to a byte that is the start of a UTF-8 character. If not, it returns
4373 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4374 this is necessary because there may be backward assertions in the pattern.
4375 When matching the same subject several times, it may save resources to use
4376 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4378 4. The code for checking the validity of UTF-8 strings has been tightened so
4379 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4380 containing "overlong sequences".
4382 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4383 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4384 should have been "&", but it just so happened that all the cases this let
4385 through by mistake were picked up later in the function.
4387 6. I had used a variable called "isblank" - this is a C99 function, causing
4388 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4390 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4391 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4392 so the version is shown; (c) stop "make test" echoing "./RunTest".
4394 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4396 9. The prototype for memmove() for systems that don't have it was using
4397 size_t, but the inclusion of the header that defines size_t was later. I've
4398 moved the #includes for the C headers earlier to avoid this.
4400 10. Added some adjustments to the code to make it easier to compiler on certain
4401 special systems:
4403 (a) Some "const" qualifiers were missing.
4404 (b) Added the macro EXPORT before all exported functions; by default this
4405 is defined to be empty.
4406 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4407 that it reads its output file name as an argument instead of writing
4408 to the standard output and assuming this can be redirected.
4410 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4411 class containing characters with values greater than 255, PCRE compilation
4412 went into a loop.
4414 12. A recursive reference to a subpattern that was within another subpattern
4415 that had a minimum quantifier of zero caused PCRE to crash. For example,
4416 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4417 recursion. If the recursively-called subpattern itself had a zero repeat,
4418 that was OK.
4420 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4421 buffer into which it was copied (for escape processing) was still set at
4422 1024, so long lines caused crashes.
4424 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4425 "internal error: code overflow...". This applied to any character class
4426 that was followed by a possessive quantifier.
4428 15. Modified the Makefile to add libpcre.la as a prerequisite for
4429 libpcreposix.la because I was told this is needed for a parallel build to
4430 work.
4432 16. If a pattern that contained .* following optional items at the start was
4433 studied, the wrong optimizing data was generated, leading to matching
4434 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4435 matching string must start with a or b or c. The correct conclusion for
4436 this pattern is that a match can start with any character.
4439 Version 4.4 13-Aug-03
4440 ---------------------
4442 1. In UTF-8 mode, a character class containing characters with values between
4443 127 and 255 was not handled correctly if the compiled pattern was studied.
4444 In fixing this, I have also improved the studying algorithm for such
4445 classes (slightly).
4447 2. Three internal functions had redundant arguments passed to them. Removal
4448 might give a very teeny performance improvement.
4450 3. Documentation bug: the value of the capture_top field in a callout is *one
4451 more than* the number of the hightest numbered captured substring.
4453 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4454 in incorrectly linking with a previously installed version. They now link
4455 explicitly with libpcre.la.
4457 5. configure.in no longer needs to recognize Cygwin specially.
4459 6. A problem in pcre.in for Windows platforms is fixed.
4461 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4462 pcretest, it used to include the size of the study block as part of its
4463 output. Unfortunately, the structure contains a field that has a different
4464 size on different hardware architectures. This meant that the tests that
4465 showed this size failed. As the block is currently always of a fixed size,
4466 this information isn't actually particularly useful in pcretest output, so
4467 I have just removed it.
4469 8. Three pre-processor statements accidentally did not start in column 1.
4470 Sadly, there are *still* compilers around that complain, even though
4471 standard C has not required this for well over a decade. Sigh.
4473 9. In pcretest, the code for checking callouts passed small integers in the
4474 callout_data field, which is a void * field. However, some picky compilers
4475 complained about the casts involved for this on 64-bit systems. Now
4476 pcretest passes the address of the small integer instead, which should get
4477 rid of the warnings.
4479 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4480 both compile and run time, and gives an error if an invalid UTF-8 sequence
4481 is found. There is a option for disabling this check in cases where the
4482 string is known to be correct and/or the maximum performance is wanted.
4484 11. In response to a bug report, I changed one line in Makefile.in from
4486 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4487 to
4488 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4490 to look similar to other lines, but I have no way of telling whether this
4491 is the right thing to do, as I do not use Windows. No doubt I'll get told
4492 if it's wrong...
4495 Version 4.3 21-May-03
4496 ---------------------
4498 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4499 Makefile.
4501 2. Some refactoring to improve the quality of the code:
4503 (i) The utf8_table... variables are now declared "const".
4505 (ii) The code for \cx, which used the "case flipping" table to upper case
4506 lower case letters, now just substracts 32. This is ASCII-specific,
4507 but the whole concept of \cx is ASCII-specific, so it seems
4508 reasonable.
4510 (iii) PCRE was using its character types table to recognize decimal and
4511 hexadecimal digits in the pattern. This is silly, because it handles
4512 only 0-9, a-f, and A-F, but the character types table is locale-
4513 specific, which means strange things might happen. A private
4514 table is now used for this - though it costs 256 bytes, a table is
4515 much faster than multiple explicit tests. Of course, the standard
4516 character types table is still used for matching digits in subject
4517 strings against \d.
4519 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4520 ending in _t are). So I've renamed it as ESC_tee.
4522 3. The first argument for regexec() in the POSIX wrapper should have been
4523 defined as "const".
4525 4. Changed pcretest to use malloc() for its buffers so that they can be
4526 Electric Fenced for debugging.
4528 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4529 to read one or more bytes before the start of the subject string. Often this
4530 had no effect on PCRE's behaviour, but in some circumstances it could
4531 provoke a segmentation fault.
4533 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4534 to try to read one or more bytes before the start of the subject string.
4536 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4537 UTF-8 support could misbehave in various ways if the subject string
4538 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4539 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4540 back over UTF-8 characters.)
4543 Version 4.2 14-Apr-03
4544 ---------------------
4546 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4548 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4549 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4550 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4551 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4552 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4554 Note: automatic setting of the BUILD variables is not yet working
4555 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4556 compile-time but not at link-time
4557 [LINK]: use for linking executables only
4558 make different versions for Windows and non-Windows
4559 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4560 libraries
4561 [LINK_FOR_BUILD]: new variable
4562 [OBJEXT]: use throughout
4563 [EXEEXT]: use throughout
4564 <winshared>: new target
4565 <wininstall>: new target
4566 <dftables.o>: use native compiler
4567 <dftables>: use native linker
4568 <install>: handle Windows platform correctly
4569 <clean>: ditto
4570 <check>: ditto
4571 copy DLL to top builddir before testing
4573 As part of these changes, -no-undefined was removed again. This was reported
4574 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4575 in any case.
4577 3. Some tidies to get rid of compiler warnings:
4579 . In the match_data structure, match_limit was an unsigned long int, whereas
4580 match_call_count was an int. I've made them both unsigned long ints.
4582 . In pcretest the fact that a const uschar * doesn't automatically cast to
4583 a void * provoked a warning.
4585 . Turning on some more compiler warnings threw up some "shadow" variables
4586 and a few more missing casts.
4588 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4589 option, a class that contained a single character with a value between 128
4590 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4592 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4593 option, a class that contained several characters, but with at least one
4594 whose value was between 128 and 255 caused PCRE to crash.
4597 Version 4.1 12-Mar-03
4598 ---------------------
4600 1. Compiling with gcc -pedantic found a couple of places where casts were
4601 needed, and a string in dftables.c that was longer than standard compilers are
4602 required to support.
4604 2. Compiling with Sun's compiler found a few more places where the code could
4605 be tidied up in order to avoid warnings.
4607 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4608 first of these names is deprecated in the latest Autoconf in favour of the name
4609 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4610 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4611 analogy I have changed it to CFLAGS_FOR_BUILD.
4613 4. Added -no-undefined to the linking command in the Makefile, because this is
4614 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4615 linking step for the pcreposix library.
4617 5. PCRE was failing to diagnose the case of two named groups with the same
4618 name.
4620 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4621 literal character that is needed in the subject for a match, and scans along to
4622 ensure that it is present before embarking on the full matching process. This
4623 saves time in cases of nested unlimited repeats that are never going to match.
4624 Problem: the scan can take a lot of time if the subject is very long (e.g.
4625 megabytes), thus penalizing straightforward matches. It is now done only if the
4626 amount of subject to be scanned is less than 1000 bytes.
4628 7. A lesser problem with the same optimization is that it was recording the
4629 first character of an anchored pattern as "needed", thus provoking a search
4630 right along the subject, even when the first match of the pattern was going to
4631 fail. The "needed" character is now not set for anchored patterns, unless it
4632 follows something in the pattern that is of non-fixed length. Thus, it still
4633 fulfils its original purpose of finding quick non-matches in cases of nested
4634 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4637 Version 4.0 17-Feb-03
4638 ---------------------
4640 1. If a comment in an extended regex that started immediately after a meta-item
4641 extended to the end of string, PCRE compiled incorrect data. This could lead to
4642 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4644 2. Moved to autoconf 2.53 and libtool 1.4.2.
4646 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4647 the special perltest8 script is no longer needed - all the tests can be run
4648 from a single perltest script.
4650 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4651 by \s. It has now been removed in PCRE. This means it isn't recognized as
4652 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4653 class [:space:] *does* include VT, thereby creating a mess.
4655 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4656 space and tab.
4658 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4659 its new features into the main test script, reducing the number of scripts.
4661 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4662 were backward compatible, and made the (?i) apply to the whole pattern, as if
4663 /i were given. Now it behaves more logically, and applies the option setting
4664 only to what follows. PCRE has been changed to follow suit. However, if it
4665 finds options settings right at the start of the pattern, it extracts them into
4666 the global options, as before. Thus, they show up in the info data.
4668 8. Added support for the \Q...\E escape sequence. Characters in between are
4669 treated as literals. This is slightly different from Perl in that $ and @ are
4670 also handled as literals inside the quotes. In Perl, they will cause variable
4671 interpolation. Note the following examples:
4673 Pattern PCRE matches Perl matches
4675 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4676 \Qabc\$xyz\E abc\$xyz abc\$xyz
4677 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4679 For compatibility with Perl, \Q...\E sequences are recognized inside character
4680 classes as well as outside them.
4682 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4683 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4684 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4685 signed/unsigned warnings.
4687 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4688 option for pcretest, so I've replaced it by a simple function that does just
4689 that job.
4691 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4692 "pcregrep -".
4694 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4695 Java package. This provides some syntactic sugar for simple cases of what my
4696 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4697 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4698 item, you can use this simplified notation. Note that only makes sense with
4699 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4700 greediness, whatever the setting of the PCRE_UNGREEDY option.
4702 13. A change of greediness default within a pattern was not taking effect at
4703 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4704 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4705 was abstracted outside.
4707 14. PCRE now supports the \G assertion. It is true when the current matching
4708 position is at the start point of the match. This differs from \A when the
4709 starting offset is non-zero. Used with the /g option of pcretest (or similar
4710 code), it works in the same way as it does for Perl's /g option. If all
4711 alternatives of a regex begin with \G, the expression is anchored to the start
4712 match position, and the "anchored" flag is set in the compiled expression.
4714 15. Some bugs concerning the handling of certain option changes within patterns
4715 have been fixed. These applied to options other than (?ims). For example,
4716 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4717 been the other way round. Some of this was related to change 7 above.
4719 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4720 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4721 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4722 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4724 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4725 mode. Unlike ".", it always matches newline, whatever the setting of
4726 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4727 assertions. Perl allows it, but it doesn't (in general) work because it can't
4728 calculate the length of the lookbehind. At least, that's the case for Perl
4729 5.8.0 - I've been told they are going to document that it doesn't work in
4730 future.
4732 18. Added an error diagnosis for escapes that PCRE does not support: these are
4733 \L, \l, \N, \P, \p, \U, \u, and \X.
4735 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4736 reading past the end of the pattern in cases such as /[abcd/.
4738 20. PCRE was getting more memory than necessary for patterns with classes that
4739 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4741 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4742 compiling PCRE for use with Virtual Pascal.
4744 22. Small fix to the Makefile to make it work properly if the build is done
4745 outside the source tree.
4747 23. Added a new extension: a condition to go with recursion. If a conditional
4748 subpattern starts with (?(R) the "true" branch is used if recursion has
4749 happened, whereas the "false" branch is used only at the top level.
4751 24. When there was a very long string of literal characters (over 255 bytes
4752 without UTF support, over 250 bytes with UTF support), the computation of how
4753 much memory was required could be incorrect, leading to segfaults or other
4754 strange effects.
4756 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4757 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4758 there was a subsequent back reference to those brackets. This meant that, for
4759 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4760 possible to check for precisely this case. All we can do is abandon the
4761 optimization if .* occurs inside capturing brackets when there are any back
4762 references whatsoever. (See below for a better fix that came later.)
4764 26. The handling of the optimization for finding the first character of a
4765 non-anchored pattern, and for finding a character that is required later in the
4766 match were failing in some cases. This didn't break the matching; it just
4767 failed to optimize when it could. The way this is done has been re-implemented.
4769 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4771 28. Added a new feature that provides some of the functionality that Perl
4772 provides with (?{...}). The facility is termed a "callout". The way it is done
4773 in PCRE is for the caller to provide an optional function, by setting
4774 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4775 global variable. By default it is unset, which disables all calling out. To get
4776 the function called, the regex must include (?C) at appropriate points. This
4777 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4778 This provides a means of identifying different callout points. When PCRE
4779 reaches such a point in the regex, if pcre_callout has been set, the external
4780 function is called. It is provided with data in a structure called
4781 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4782 matching continues; if it returns a non-zero value, the match at the current
4783 point fails. However, backtracking will occur if possible. [This was changed
4784 later and other features added - see item 49 below.]
4786 29. pcretest is upgraded to test the callout functionality. It provides a
4787 callout function that displays information. By default, it shows the start of
4788 the match and the current position in the text. There are some new data escapes
4789 to vary what happens:
4791 \C+ in addition, show current contents of captured substrings
4792 \C- do not supply a callout function
4793 \C!n return 1 when callout number n is reached
4794 \C!n!m return 1 when callout number n is reached for the mth time
4796 30. If pcregrep was called with the -l option and just a single file name, it
4797 output "<stdin>" if a match was found, instead of the file name.
4799 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4800 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4801 pcre_exec(). This saves a malloc/free per call. The default value of
4802 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4803 when configuring.
4805 32. The default maximum size of a compiled pattern is 64K. There have been a
4806 few cases of people hitting this limit. The code now uses macros to handle the
4807 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4808 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4809 configuring. Tests 2 and 5 work only with 2-byte links because they output
4810 debugging information about compiled patterns.
4812 33. Internal code re-arrangements:
4814 (a) Moved the debugging function for printing out a compiled regex into
4815 its own source file (printint.c) and used #include to pull it into
4816 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4817 separate copies.
4819 (b) Defined the list of op-code names for debugging as a macro in
4820 internal.h so that it is next to the definition of the opcodes.
4822 (c) Defined a table of op-code lengths for simpler skipping along compiled
4823 code. This is again a macro in internal.h so that it is next to the
4824 definition of the opcodes.
4826 34. Added support for recursive calls to individual subpatterns, along the
4827 lines of Robin Houston's patch (but implemented somewhat differently).
4829 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4830 allow it to read and process whole directories in Win32. This code was
4831 contributed by Lionel Fourquaux; it has not been tested by me.
4833 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4834 used to name a group. Names consist of alphanumerics and underscores, and must
4835 be unique. Back references use the syntax (?P=name) and recursive calls use
4836 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4837 numbers. The function pcre_fullinfo() can be used after compilation to extract
4838 a name/number map. There are three relevant calls:
4840 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4841 PCRE_INFO_NAMECOUNT yields the number of entries
4842 PCRE_INFO_NAMETABLE yields a pointer to the map.
4844 The map is a vector of fixed-size entries. The size of each entry depends on
4845 the length of the longest name used. The first two bytes of each entry are the
4846 group number, most significant byte first. There follows the corresponding
4847 name, zero terminated. The names are in alphabetical order.
4849 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4850 case instead of 255. Making it the same both with and without UTF-8 support
4851 means that the same test output works with both.
4853 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4854 calling malloc() with a zero argument.
4856 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4857 optimization. I've improved things by keeping a bitmap of backreferences with
4858 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4859 fact referenced, the optimization can be applied. It is unlikely that a
4860 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4861 the match to follow \n) will appear inside brackets with a number greater than
4862 31, but if it does, any back reference > 31 suppresses the optimization.
4864 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4865 of disabling numbered capturing parentheses. Any opening parenthesis that is
4866 not followed by ? behaves as if it were followed by ?: but named parentheses
4867 can still be used for capturing (and they will acquire numbers in the usual
4868 way).
4870 41. Redesigned the return codes from the match() function into yes/no/error so
4871 that errors can be passed back from deep inside the nested calls. A malloc
4872 failure while inside a recursive subpattern call now causes the
4873 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4875 42. It is now possible to set a limit on the number of times the match()
4876 function is called in a call to pcre_exec(). This facility makes it possible to
4877 limit the amount of recursion and backtracking, though not in a directly
4878 obvious way, because the match() function is used in a number of different
4879 circumstances. The count starts from zero for each position in the subject
4880 string (for non-anchored patterns). The default limit is, for compatibility, a
4881 large number, namely 10 000 000. You can change this in two ways:
4883 (a) When configuring PCRE before making, you can use --with-match-limit=n
4884 to set a default value for the compiled library.
4886 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4887 a different value is set. See 45 below.
4889 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4891 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4892 of things that can be changed at compile time. The first argument specifies
4893 what is wanted and the second points to where the information is to be placed.
4894 The current list of available information is:
4898 The output is an integer that is set to one if UTF-8 support is available;
4899 otherwise it is set to zero.
4903 The output is an integer that it set to the value of the code that is used for
4904 newline. It is either LF (10) or CR (13).
4908 The output is an integer that contains the number of bytes used for internal
4909 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4913 The output is an integer that contains the threshold above which the POSIX
4914 interface uses malloc() for output vectors. See item 31 above.
4918 The output is an unsigned integer that contains the default limit of the number
4919 of match() calls in a pcre_exec() execution. See 42 above.
4921 44. pcretest has been upgraded by the addition of the -C option. This causes it
4922 to extract all the available output from the new pcre_config() function, and to
4923 output it. The program then exits immediately.
4925 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4926 order to support additional features. One way would have been to define
4927 pcre_exec2() (for example) with extra arguments, but this would not have been
4928 extensible, and would also have required all calls to the original function to
4929 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4930 is used for passing in "extra" data from pcre_study().
4932 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4933 contains the following fields:
4935 flags a bitmap indicating which of the following fields are set
4936 study_data opaque data from pcre_study()
4937 match_limit a way of specifying a limit on match() calls for a specific
4938 call to pcre_exec()
4939 callout_data data for callouts (see 49 below)
4941 The flag bits are also defined in pcre.h, and are
4947 The pcre_study() function now returns one of these new pcre_extra blocks, with
4948 the actual study data pointed to by the study_data field, and the
4949 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4950 before. That is, this change is entirely upwards-compatible and requires no
4951 change to existing code.
4953 If you want to pass in additional data to pcre_exec(), you can either place it
4954 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4955 block.
4957 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4958 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4959 times with different match limits, until it finds the minimum value needed for
4960 pcre_exec() to complete. The value is then output. This can be instructive; for
4961 most simple matches the number is quite small, but for pathological cases it
4962 gets very large very quickly.
4964 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4965 returns the size of the data block pointed to by the study_data field in a
4966 pcre_extra block, that is, the value that was passed as the argument to
4967 pcre_malloc() when PCRE was getting memory in which to place the information
4968 created by pcre_study(). The fourth argument should point to a size_t variable.
4969 pcretest has been extended so that this information is shown after a successful
4970 pcre_study() call when information about the compiled regex is being displayed.
4972 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4973 because what follows is always an absolute path. (Later: it turns out that this
4974 is more than cosmetic for MinGW, because it doesn't like empty path
4975 components.)
4977 49. Some changes have been made to the callout feature (see 28 above):
4979 (i) A callout function now has three choices for what it returns:
4981 0 => success, carry on matching
4982 > 0 => failure at this point, but backtrack if possible
4983 < 0 => serious error, return this value from pcre_exec()
4985 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4986 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4987 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4988 use by callout functions. It will never be used by PCRE itself.
4990 (ii) The pcre_extra structure (see 45 above) has a void * field called
4991 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4992 pcre_callout_block structure has a field of the same name. The contents of
4993 the field passed in the pcre_extra structure are passed to the callout
4994 function in the corresponding field in the callout block. This makes it
4995 easier to use the same callout-containing regex from multiple threads. For
4996 testing, the pcretest program has a new data escape
4998 \C*n pass the number n (may be negative) as callout_data
5000 If the callout function in pcretest receives a non-zero value as
5001 callout_data, it returns that value.
5003 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5004 there were some redundant $(CFLAGS) in commands that are now specified as
5005 $(LINK), which already includes $(CFLAGS).
5007 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5008 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5009 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5010 one-byte characters throughout. Note that case-insensitive matching applies
5011 only to characters whose values are less than 256. PCRE doesn't support the
5012 notion of cases for higher-valued characters.
5014 (i) A character class whose characters are all within 0-255 is handled as
5015 a bit map, and the map is inverted for negative classes. Previously, a
5016 character > 255 always failed to match such a class; however it should
5017 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5019 (ii) A negated character class with a single character < 255 is coded as
5020 "not this character" (OP_NOT). This wasn't working properly when the test
5021 character was multibyte, either singly or repeated.
5023 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5024 mode, for example: \x{100}{2,3}.
5026 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5027 singly or repeated) now correctly test multibyte characters. However,
5028 PCRE doesn't recognize any characters with values greater than 255 as
5029 digits, spaces, or word characters. Such characters always match \D, \S,
5030 and \W, and never match \d, \s, or \w.
5032 (v) Classes may now contain characters and character ranges with values
5033 greater than 255. For example: [ab\x{100}-\x{400}].
5035 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5036 PCRE in UTF-8 mode.
5038 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5039 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5040 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5041 value.)
5043 53. The single man page has become too large. I have therefore split it up into
5044 a number of separate man pages. These also give rise to individual HTML pages;
5045 these are now put in a separate directory, and there is an index.html page that
5046 lists them all. Some hyperlinking between the pages has been installed.
5048 54. Added convenience functions for handling named capturing parentheses.
5050 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5051 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5052 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5053 are faulted.
5055 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5056 calling configure. These values are used when compiling the dftables.c program
5057 which is run to generate the source of the default character tables. They
5058 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5059 you will need to set these values.
5061 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5064 Version 3.9 02-Jan-02
5065 ---------------------
5067 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5069 2. If --disable-static was given, the building process failed when trying to
5070 build pcretest and pcregrep. (For some reason it was using libtool to compile
5071 them, which is not right, as they aren't part of the library.)
5074 Version 3.8 18-Dec-01
5075 ---------------------
5077 1. The experimental UTF-8 code was completely screwed up. It was packing the
5078 bytes in the wrong order. How dumb can you get?
5081 Version 3.7 29-Oct-01
5082 ---------------------
5084 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5085 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5086 this didn't happen under Solaris 8, where I normally test things.
5088 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5089 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5090 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5091 this fix an hour or so after the initial 3.7 release.)
5094 Version 3.6 23-Oct-01
5095 ---------------------
5097 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5098 offsets passed as NULL with zero offset count.
5100 2. The config.guess and config.sub files had not been updated when I moved to
5101 the latest autoconf.
5104 Version 3.5 15-Aug-01
5105 ---------------------
5107 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5108 had been forgotten.
5110 2. By using declared but undefined structures, we can avoid using "void"
5111 definitions in pcre.h while keeping the internal definitions of the structures
5112 private.
5114 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5115 user point of view, this means that both static and shared libraries are built
5116 by default, but this can be individually controlled. More of the work of
5117 handling this static/shared cases is now inside libtool instead of PCRE's make
5118 file.
5120 4. The pcretest utility is now installed along with pcregrep because it is
5121 useful for users (to test regexs) and by doing this, it automatically gets
5122 relinked by libtool. The documentation has been turned into a man page, so
5123 there are now .1, .txt, and .html versions in /doc.
5125 5. Upgrades to pcregrep:
5126 (i) Added long-form option names like gnu grep.
5127 (ii) Added --help to list all options with an explanatory phrase.
5128 (iii) Added -r, --recursive to recurse into sub-directories.
5129 (iv) Added -f, --file to read patterns from a file.
5131 6. pcre_exec() was referring to its "code" argument before testing that
5132 argument for NULL (and giving an error if it was NULL).
5134 7. Upgraded Makefile.in to allow for compiling in a different directory from
5135 the source directory.
5137 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5138 options bits, the pointer it was passed was to an int instead of to an unsigned
5139 long int. This mattered only on 64-bit systems.
5141 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5142 generated) instead of pcre.in, which it its source. Also made the same change
5143 in several of the .c files.
5145 10. A new release of gcc defines printf() as a macro, which broke pcretest
5146 because it had an ifdef in the middle of a string argument for printf(). Fixed
5147 by using separate calls to printf().
5149 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5150 script, to force use of CR or LF instead of \n in the source. On non-Unix
5151 systems, the value can be set in config.h.
5153 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5154 absolute limit. Changed the text of the error message to make this clear, and
5155 likewise updated the man page.
5157 13. The limit of 99 on the number of capturing subpatterns has been removed.
5158 The new limit is 65535, which I hope will not be a "real" limit.
5161 Version 3.4 22-Aug-00
5162 ---------------------
5164 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5166 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5169 Version 3.3 01-Aug-00
5170 ---------------------
5172 1. If an octal character was given, but the value was greater than \377, it
5173 was not getting masked to the least significant bits, as documented. This could
5174 lead to crashes in some systems.
5176 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5177 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5179 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5180 These just pass their arguments on to (pcre_free)(), but they are provided
5181 because some uses of PCRE bind it to non-C systems that can call its functions,
5182 but cannot call free() or pcre_free() directly.
5184 4. Add "make test" as a synonym for "make check". Corrected some comments in
5185 the Makefile.
5187 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5188 Makefile.
5190 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5191 command called pgrep for grepping around the active processes.
5193 7. Added the beginnings of support for UTF-8 character strings.
5195 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5196 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5197 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5198 out for the ar command.)
5201 Version 3.2 12-May-00
5202 ---------------------
5204 This is purely a bug fixing release.
5206 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5207 of ZA. This was just one example of several cases that could provoke this bug,
5208 which was introduced by change 9 of version 2.00. The code for breaking
5209 infinite loops after an iteration that matches an empty string was't working
5210 correctly.
5212 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5213 when matched against abbab (for example). After matching an empty string, it
5214 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5215 caused it to match further down the string than it should.
5217 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5218 was there because it doesn't seem to be needed, and it causes trouble on some
5219 systems, as it is not a Standard C header. It has been removed.
5221 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5222 were reported on the Macintosh. The changes were from
5224 while ((c = *(++ptr)) != 0 && c != '\n');
5225 to
5226 while ((c = *(++ptr)) != 0 && c != '\n') ;
5228 Totally extraordinary, but if that's what it takes...
5230 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5231 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5232 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5233 assumes the way PCRE uses memmove() (always moving upwards).
5235 6. PCRE is being used in one environment where strchr() is not available. There
5236 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5237 faster code anyway.
5240 Version 3.1 09-Feb-00
5241 ---------------------
5243 The only change in this release is the fixing of some bugs in Makefile.in for
5244 the "install" target:
5246 (1) It was failing to install pcreposix.h.
5248 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5251 Version 3.0 01-Feb-00
5252 ---------------------
5254 1. Add support for the /+ modifier to perltest (to output $` like it does in
5255 pcretest).
5257 2. Add support for the /g modifier to perltest.
5259 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5260 matches null strings.
5262 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5263 pattern. Perl treats empty patterns specially - it reuses the most recent
5264 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5265 effect.
5267 5. The POSIX interface was broken in that it was just handing over the POSIX
5268 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5269 required a bigger vector, with some working space on the end. This means that
5270 the POSIX wrapper now has to get and free some memory, and copy the results.
5272 6. Added some simple autoconf support, placing the test data and the
5273 documentation in separate directories, re-organizing some of the
5274 information files, and making it build pcre-config (a GNU standard). Also added
5275 libtool support for building PCRE as a shared library, which is now the
5276 default.
5278 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5279 09 are not valid octal constants. Single digits will be used for minor values
5280 less than 10.
5282 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5283 existing programs that set these in the POSIX interface can use PCRE without
5284 modification.
5286 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5287 return all that pcre_info() returns, plus additional data. The pcre_info()
5288 function is retained for compatibility, but is considered to be obsolete.
5290 10. Added experimental recursion feature (?R) to handle one common case that
5291 Perl 5.6 will be able to do with (?p{...}).
5293 11. Added support for POSIX character classes like [:alpha:], which Perl is
5294 adopting.
5297 Version 2.08 31-Aug-99
5298 ----------------------
5300 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5301 trying to match at the startoffset position, but instead was moving forward to
5302 the next newline as if a previous match had failed.
5304 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5305 and could get into a loop if a null string was matched other than at the start
5306 of the subject.
5308 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5309 be distinguished at compile time, and for completeness also added PCRE_DATE.
5311 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5312 in GnuWin32 environments.
5315 Version 2.07 29-Jul-99
5316 ----------------------
5318 1. The documentation is now supplied in plain text form and HTML as well as in
5319 the form of man page sources.
5321 2. C++ compilers don't like assigning (void *) values to other pointer types.
5322 In particular this affects malloc(). Although there is no problem in Standard
5323 C, I've put in casts to keep C++ compilers happy.
5325 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5326 should be (const char *).
5328 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5329 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5330 However, I haven't made this a standard facility. The documentation doesn't
5331 mention it, and the Makefile doesn't support it.
5333 5. The Makefile now contains an "install" target, with editable destinations at
5334 the top of the file. The pcretest program is not installed.
5336 6. pgrep -V now gives the PCRE version number and date.
5338 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5339 causing the entire string to be ignored, instead of just the last character.
5341 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5342 non-matching string, it can take a very, very long time, even for strings of
5343 quite modest length, because of the nested recursion. PCRE now does better in
5344 some of these cases. It does this by remembering the last required literal
5345 character in the pattern, and pre-searching the subject to ensure it is present
5346 before running the real match. In other words, it applies a heuristic to detect
5347 some types of certain failure quickly, and in the above example, if presented
5348 with a string that has no trailing " it gives "no match" very quickly.
5350 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5351 other alternatives are tried instead.
5354 Version 2.06 09-Jun-99
5355 ----------------------
5357 1. Change pcretest's output for amount of store used to show just the code
5358 space, because the remainder (the data block) varies in size between 32-bit and
5359 64-bit systems.
5361 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5362 start matching at. This allows lookbehinds to work when searching for multiple
5363 occurrences in a string.
5365 3. Added additional options to pcretest for testing multiple occurrences:
5367 /+ outputs the rest of the string that follows a match
5368 /g loops for multiple occurrences, using the new startoffset argument
5369 /G loops for multiple occurrences by passing an incremented pointer
5371 4. PCRE wasn't doing the "first character" optimization for patterns starting
5372 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5373 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5374 the letter 'x'. On long subject strings, this gives a significant speed-up.
5377 Version 2.05 21-Apr-99
5378 ----------------------
5380 1. Changed the type of magic_number from int to long int so that it works
5381 properly on 16-bit systems.
5383 2. Fixed a bug which caused patterns starting with .* not to work correctly
5384 when the subject string contained newline characters. PCRE was assuming
5385 anchoring for such patterns in all cases, which is not correct because .* will
5386 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5387 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5388 must be retried after every newline in the subject.
5391 Version 2.04 18-Feb-99
5392 ----------------------
5394 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5395 computation of the store needed to hold the pattern was incorrect (too large).
5396 If such patterns were nested a few deep, this could multiply and become a real
5397 problem.
5399 2. Added /M option to pcretest to show the memory requirement of a specific
5400 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5402 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5403 compiled in such a way that the backtracking after subsequent failure was
5404 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5405 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5408 Version 2.03 02-Feb-99
5409 ----------------------
5411 1. Fixed typo and small mistake in man page.
5413 2. Added 4th condition (GPL supersedes if conflict) and created separate
5414 LICENCE file containing the conditions.
5416 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5417 Perl, that is the internal \ allows the delimiter to be included in the
5418 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5419 the final delimiter, add \ to the end of the pattern (to test the error).
5421 4. Added the convenience functions for extracting substrings after a successful
5422 match. Updated pcretest to make it able to test these functions.
5425 Version 2.02 14-Jan-99
5426 ----------------------
5428 1. Initialized the working variables associated with each extraction so that
5429 their saving and restoring doesn't refer to uninitialized store.
5431 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5432 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5433 fix the problem.
5435 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5436 calls, and wasn't printing the correct value for compiling calls. Increased the
5437 default value of LOOPREPEAT, and the number of significant figures in the
5438 times.
5440 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5442 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5443 a building problem on Windows NT with a FAT file system.
5446 Version 2.01 21-Oct-98
5447 ----------------------
5449 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5450 to character tables built by pcre_maketables() in the current locale. If NULL
5451 is passed, the default tables are used.
5454 Version 2.00 24-Sep-98
5455 ----------------------
5457 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5458 it any more.
5460 2. Allow quantification of (?>) groups, and make it work correctly.
5462 3. The first character computation wasn't working for (?>) groups.
5464 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5465 end of the subject) and add 5.005's \z, which really does match only at the
5466 very end of the subject.
5468 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5470 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5471 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5472 localized options. All options to pcre_study() were also removed.
5474 7. Add other new features from 5.005:
5476 $(?<= positive lookbehind
5477 $(?<! negative lookbehind
5478 (?imsx-imsx) added the unsetting capability
5479 such a setting is global if at outer level; local otherwise
5480 (?imsx-imsx:) non-capturing groups with option setting
5481 (?(cond)re|re) conditional pattern matching
5483 A backreference to itself in a repeated group matches the previous
5484 captured string.
5486 8. General tidying up of studying (both automatic and via "study")
5487 consequential on the addition of new assertions.
5489 9. As in 5.005, unlimited repeated groups that could match an empty substring
5490 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5491 runtime if any iteration does actually match an empty substring.
5493 10. Include the RunTest script in the distribution.
5495 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5496 discrepancies, some of which were old and were also with respect to 5.004. They
5497 have now been fixed.
5500 Version 1.09 28-Apr-98
5501 ----------------------
5503 1. A negated single character class followed by a quantifier with a minimum
5504 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5505 program crashes, or just wrong answers. This did not apply to negated classes
5506 containing more than one character, or to minima other than one.
5509 Version 1.08 27-Mar-98
5510 ----------------------
5512 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5514 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5515 latter must appear before anything that relies on it in the pattern.
5518 Version 1.07 16-Feb-98
5519 ----------------------
5521 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5522 repeat of a potentially empty string).
5525 Version 1.06 23-Jan-98
5526 ----------------------
5528 1. Added Markus Oberhumer's little patches for C++.
5530 2. Literal strings longer than 255 characters were broken.
5533 Version 1.05 23-Dec-97
5534 ----------------------
5536 1. Negated character classes containing more than one character were failing if
5537 PCRE_CASELESS was set at run time.
5540 Version 1.04 19-Dec-97
5541 ----------------------
5543 1. Corrected the man page, where some "const" qualifiers had been omitted.
5545 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5546 input syntax.
5548 3. Fixed memory leak which occurred when a regex with back references was
5549 matched with an offsets vector that wasn't big enough. The temporary memory
5550 that is used in this case wasn't being freed if the match failed.
5552 4. Tidied pcretest to ensure it frees memory that it gets.
5554 5. Temporary memory was being obtained in the case where the passed offsets
5555 vector was exactly big enough.
5557 6. Corrected definition of offsetof() from change 5 below.
5559 7. I had screwed up change 6 below and broken the rules for the use of
5560 setjmp(). Now fixed.
5563 Version 1.03 18-Dec-97
5564 ----------------------
5566 1. A erroneous regex with a missing opening parenthesis was correctly
5567 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5568 on some systems.
5570 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5571 it was reported that one broken compiler failed on the former because "code" is
5572 also an independent variable.
5574 3. The erroneous regex a[]b caused an array overrun reference.
5576 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5577 fail on data ending with that character. (It was going on too far, and checking
5578 the next character, typically a binary zero.) This was specific to the
5579 optimized code for single-character negative classes.
5581 5. Added a contributed patch from the TIN world which does the following:
5583 + Add an undef for memmove, in case the the system defines a macro for it.
5585 + Add a definition of offsetof(), in case there isn't one. (I don't know
5586 the reason behind this - offsetof() is part of the ANSI standard - but
5587 it does no harm).
5589 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5590 most of the places where whitespace preceded '#'. I have given up and
5591 allowed the remaining 2 cases to be at the margin.
5593 + Rename some variables in pcre to eliminate shadowing. This seems very
5594 pedantic, but does no harm, of course.
5596 6. Moved the call to setjmp() into its own function, to get rid of warnings
5597 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5599 7. Constructs such as \d{8,} were compiling into the equivalent of
5600 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5601 outcome, but in this particular case used more store than had been allocated,
5602 which caused the bug to be discovered because it threw up an internal error.
5604 8. The debugging code in both pcre and pcretest for outputting the compiled
5605 form of a regex was going wrong in the case of back references followed by
5606 curly-bracketed repeats.
5609 Version 1.02 12-Dec-97
5610 ----------------------
5612 1. Typos in pcre.3 and comments in the source fixed.
5614 2. Applied a contributed patch to get rid of places where it used to remove
5615 'const' from variables, and fixed some signed/unsigned and uninitialized
5616 variable warnings.
5618 3. Added the "runtest" target to Makefile.
5620 4. Set default compiler flag to -O2 rather than just -O.
5623 Version 1.01 19-Nov-97
5624 ----------------------
5626 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5627 like /([ab]*)*/, that is, for classes with more than one character in them.
5629 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5630 as /((?>a*))*/ (a PCRE_EXTRA facility).
5633 Version 1.00 18-Nov-97
5634 ----------------------
5636 1. Added compile-time macros to support systems such as SunOS4 which don't have
5637 memmove() or strerror() but have other things that can be used instead.
5639 2. Arranged that "make clean" removes the executables.
5642 Version 0.99 27-Oct-97
5643 ----------------------
5645 1. Fixed bug in code for optimizing classes with only one character. It was
5646 initializing a 32-byte map regardless, which could cause it to run off the end
5647 of the memory it had got.
5649 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5652 Version 0.98 22-Oct-97
5653 ----------------------
5655 1. Fixed bug in code for handling temporary memory usage when there are more
5656 back references than supplied space in the ovector. This could cause segfaults.
5659 Version 0.97 21-Oct-97
5660 ----------------------
5662 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5664 2. Optimized negated single characters not to use a bit map.
5666 3. Brought error texts together as macro definitions; clarified some of them;
5667 fixed one that was wrong - it said "range out of order" when it meant "invalid
5668 escape sequence".
5670 4. Changed some char * arguments to const char *.
5672 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5674 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5675 pcretest.
5678 Version 0.96 16-Oct-97
5679 ----------------------
5681 1. Added a simple "pgrep" utility to the distribution.
5683 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5684 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5685 where "ddd" means "one or more decimal digits".
5687 3. Fixed serious bug. If a pattern had a back reference, but the call to
5688 pcre_exec() didn't supply a large enough ovector to record the related
5689 identifying subpattern, the match always failed. PCRE now remembers the number
5690 of the largest back reference, and gets some temporary memory in which to save
5691 the offsets during matching if necessary, in order to ensure that
5692 backreferences always work.
5694 4. Increased the compatibility with Perl in a number of ways:
5696 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5697 to request this handling. The option can be set at compile or exec time.
5699 (b) $ matches before a terminating newline by default; an option
5700 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5701 mode). The option can be set at compile or exec time.
5703 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5704 the same as Perl's. If the decimal number it represents is less than 10
5705 or there aren't that many previous left capturing parentheses, an octal
5706 escape is read. Inside a character class, it's always an octal escape,
5707 even if it is a single digit.
5709 (d) An escaped but undefined alphabetic character is taken as a literal,
5710 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5711 escapes.
5713 (e) {0} is now permitted. (The previous item is removed from the compiled
5714 pattern).
5716 5. Changed all the names of code files so that the basic parts are no longer
5717 than 10 characters, and abolished the teeny "globals.c" file.
5719 6. Changed the handling of character classes; they are now done with a 32-byte
5720 bit map always.
5722 7. Added the -d and /D options to pcretest to make it possible to look at the
5723 internals of compilation without having to recompile pcre.
5726 Version 0.95 23-Sep-97
5727 ----------------------
5729 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5730 \x20 at the start of a run of normal characters. These were being treated as
5731 real characters, instead of the source characters being re-checked.
5734 Version 0.94 18-Sep-97
5735 ----------------------
5737 1. The functions are now thread-safe, with the caveat that the global variables
5738 containing pointers to malloc() and free() or alternative functions are the
5739 same for all threads.
5741 2. Get pcre_study() to generate a bitmap of initial characters for non-
5742 anchored patterns when this is possible, and use it if passed to pcre_exec().
5745 Version 0.93 15-Sep-97
5746 ----------------------
5748 1. /(b)|(:+)/ was computing an incorrect first character.
5750 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5751 but not actually doing anything yet.
5753 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5754 as Perl does (e.g. [-az] or [az-]).
5756 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5757 all possible positions.
5759 5. Split up into different modules to avoid including unneeded functions in a
5760 compiled binary. However, compile and exec are still in one module. The "study"
5761 function is split off.
5763 6. The character tables are now in a separate module whose source is generated
5764 by an auxiliary program - but can then be edited by hand if required. There are
5765 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5766 toupper() in the code.
5768 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5769 make them global. Abolish the function for setting them, as the caller can now
5770 set them directly.
5773 Version 0.92 11-Sep-97
5774 ----------------------
5776 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5777 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5779 2. Caseless matching was not working in character classes if the characters in
5780 the pattern were in upper case.
5782 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5784 4. Make PCRE_ANCHORED public and accept as a compile option.
5786 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5787 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5788 pass them.
5790 6. Give an error if bad option bits passed at compile or run time.
5792 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5793 pcretest to cause it to pass that flag.
5795 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5796 options, and the first character, if set.
5798 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5801 Version 0.91 10-Sep-97
5802 ----------------------
5804 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5805 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5807 2. PCRE was looping on encountering an indefinitely repeated back reference to
5808 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5809 Perl does - treats the match as successful.
5811 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5