ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1600 - (show annotations)
Wed Sep 2 08:54:57 2015 UTC (4 years, 4 months ago) by ph10
File size: 284592 byte(s)
Remove use of /dev/null in test 2 because Windows doesn't like it.
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.
72 17. \h within a class was incorrectly compiled in EBCDIC environments.
74 18. A pattern with an unmatched closing parenthesis that contained a backward
75 assertion which itself contained a forward reference caused buffer
76 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
78 19. JIT should return with error when the compiled pattern requires more stack
79 space than the maximum.
81 20. A possessively repeated conditional group that could match an empty string,
82 for example, /(?(R))*+/, was incorrectly compiled.
84 21. Fix infinite recursion in the JIT compiler when certain patterns such as
85 /(?:|a|){100}x/ are analysed.
87 22. Some patterns with character classes involving [: and \\ were incorrectly
88 compiled and could cause reading from uninitialized memory or an incorrect
89 error diagnosis.
91 23. Pathological patterns containing many nested occurrences of [: caused
92 pcre_compile() to run for a very long time.
94 24. A conditional group with only one branch has an implicit empty alternative
95 branch and must therefore be treated as potentially matching an empty
96 string.
98 25. If (?R was followed by - or + incorrect behaviour happened instead of a
99 diagnostic.
101 26. Arrange to give up on finding the minimum matching length for overly
102 complex patterns.
104 27. Similar to (4) above: in a pattern with duplicated named groups and an
105 occurrence of (?| it is possible for an apparently non-recursive back
106 reference to become recursive if a later named group with the relevant
107 number is encountered. This could lead to a buffer overflow. Wen Guanxing
108 from Venustech ADLAB discovered this bug.
110 28. If pcregrep was given the -q option with -c or -l, or when handling a
111 binary file, it incorrectly wrote output to stdout.
113 29. The JIT compiler did not restore the control verb head in case of *THEN
114 control verbs. This issue was found by Karl Skomski with a custom LLVM
115 fuzzer.
117 30. Error messages for syntax errors following \g and \k were giving inaccurate
118 offsets in the pattern.
120 31. Added a check for integer overflow in conditions (?(<digits>) and
121 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
122 fuzzer.
124 32. Handling recursive references such as (?2) when the reference is to a group
125 later in the pattern uses code that is very hacked about and error-prone.
126 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
127 give an internal error if it is obvious that compiling has gone wrong.
129 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
130 This issue was found by Karl Skomski with a custom LLVM fuzzer.
132 34. The JIT compiler should restore the control chain for empty possessive
133 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
135 35. Match limit check added to JIT recursion. This issue was found by Karl
136 Skomski with a custom LLVM fuzzer.
138 36. Yet another case similar to 27 above has been circumvented by an
139 unconditional allocation of extra memory. This issue is fixed "properly" in
140 PCRE2 by refactoring the way references are handled. Wen Guanxing
141 from Venustech ADLAB discovered this bug.
143 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
144 with a custom LLVM fuzzer.
146 38. Fixed a corner case of range optimization in JIT.
148 39. An incorrect error "overran compiling workspace" was given if there were
149 exactly enough group forward references such that the last one extended
150 into the workspace safety margin. The next one would have expanded the
151 workspace. The test for overflow was not including the safety margin.
153 40. A match limit issue is fixed in JIT which was found by Karl Skomski
154 with a custom LLVM fuzzer.
156 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
157 work under Windows. (Why has it taken so long for anyone to notice?)
160 Version 8.37 28-April-2015
161 --------------------------
163 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
164 for those parentheses to be closed with whatever has been captured so far.
165 However, it was failing to mark any other groups between the hightest
166 capture so far and the currrent group as "unset". Thus, the ovector for
167 those groups contained whatever was previously there. An example is the
168 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
170 2. If an assertion condition was quantified with a minimum of zero (an odd
171 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
173 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
174 unrecognized modifier, a crash could occur.
176 4. An attempt to do global matching in pcretest with a zero-length ovector
177 caused a crash.
179 5. Fixed a memory leak during matching that could occur for a subpattern
180 subroutine call (recursive or otherwise) if the number of captured groups
181 that had to be saved was greater than ten.
183 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
184 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
185 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
187 7. A UTF pattern containing a "not" match of a non-ASCII character and a
188 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
190 8. When a pattern is compiled, it remembers the highest back reference so that
191 when matching, if the ovector is too small, extra memory can be obtained to
192 use instead. A conditional subpattern whose condition is a check on a
193 capture having happened, such as, for example in the pattern
194 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
195 setting the highest backreference number. This mattered only if pcre_exec()
196 was called with an ovector that was too small to hold the capture, and there
197 was no other kind of back reference (a situation which is probably quite
198 rare). The effect of the bug was that the condition was always treated as
199 FALSE when the capture could not be consulted, leading to a incorrect
200 behaviour by pcre_exec(). This bug has been fixed.
202 9. A reference to a duplicated named group (either a back reference or a test
203 for being set in a conditional) that occurred in a part of the pattern where
204 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
205 to be incorrectly calculated, leading to overwriting.
207 10. A mutually recursive set of back references such as (\2)(\1) caused a
208 segfault at study time (while trying to find the minimum matching length).
209 The infinite loop is now broken (with the minimum length unset, that is,
210 zero).
212 11. If an assertion that was used as a condition was quantified with a minimum
213 of zero, matching went wrong. In particular, if the whole group had
214 unlimited repetition and could match an empty string, a segfault was
215 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
216 assertions to be quantified, but not if they are being used as conditions,
217 so the above pattern is faulted by Perl. PCRE has now been changed so that
218 it also rejects such patterns.
220 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
221 failed to allow the zero-repeat case if pcre2_exec() was called with an
222 ovector too small to capture the group.
224 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
225 Red Hat Product Security:
227 (a) A crash if /K and /F were both set with the option to save the compiled
228 pattern.
230 (b) Another crash if the option to print captured substrings in a callout
231 was combined with setting a null ovector, for example \O\C+ as a subject
232 string.
234 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
235 forward reference repeated a large (but limited) number of times within a
236 repeated outer group that has a zero minimum quantifier, caused incorrect
237 code to be compiled, leading to the error "internal error:
238 previously-checked referenced subpattern not found" when an incorrect
239 memory address was read. This bug was reported as "heap overflow",
240 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
241 CVE-2015-2325.
243 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
244 call within a group that also contained a recursive back reference caused
245 incorrect code to be compiled. This bug was reported as "heap overflow",
246 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
247 number CVE-2015-2326.
249 24. Computing the size of the JIT read-only data in advance has been a source
250 of various issues, and new ones are still appear unfortunately. To fix
251 existing and future issues, size computation is eliminated from the code,
252 and replaced by on-demand memory allocation.
254 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
255 adjacent to the end of the range, and the range contained characters with
256 more than one other case, caused incorrect behaviour when compiled in UTF
257 mode. In that example, the range a-j was left out of the class.
259 26. Fix JIT compilation of conditional blocks, which assertion
260 is converted to (*FAIL). E.g: /(?(?!))/.
262 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
263 discovered by the LLVM fuzzer.
265 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
266 when this assertion was used as a condition, for example (?(?!)a|b). In
267 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
268 error about an unsupported item.
270 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
271 possessification code could take exponential time to complete. A recursion
272 depth limit of 1000 has been imposed to limit the resources used by this
273 optimization.
275 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
276 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
277 because \S ensures they are all in the class. The code for doing this was
278 interacting badly with the code for computing the amount of space needed to
279 compile the pattern, leading to a buffer overflow. This bug was discovered
280 by the LLVM fuzzer.
282 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
283 other kinds of group caused stack overflow at compile time. This bug was
284 discovered by the LLVM fuzzer.
286 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
287 between a subroutine call and its quantifier was incorrectly compiled,
288 leading to buffer overflow or other errors. This bug was discovered by the
289 LLVM fuzzer.
291 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
292 assertion after (?(. The code was failing to check the character after
293 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
294 was discovered by the LLVM fuzzer.
296 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
297 a fixed maximum following a group that contains a subroutine reference was
298 incorrectly compiled and could trigger buffer overflow. This bug was
299 discovered by the LLVM fuzzer.
301 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
302 caused a stack overflow instead of the diagnosis of a non-fixed length
303 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
305 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
306 (e.g. /(?<=\Ka)/) could make pcregrep loop.
308 37. There was a similar problem to 36 in pcretest for global matches.
310 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
311 and a subsequent item in the pattern caused a non-match, backtracking over
312 the repeated \X did not stop, but carried on past the start of the subject,
313 causing reference to random memory and/or a segfault. There were also some
314 other cases where backtracking after \C could crash. This set of bugs was
315 discovered by the LLVM fuzzer.
317 39. The function for finding the minimum length of a matching string could take
318 a very long time if mutual recursion was present many times in a pattern,
319 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
320 method has been implemented. This infelicity was discovered by the LLVM
321 fuzzer.
323 40. Static linking against the PCRE library using the pkg-config module was
324 failing on missing pthread symbols.
327 Version 8.36 26-September-2014
328 ------------------------------
330 1. Got rid of some compiler warnings in the C++ modules that were shown up by
331 -Wmissing-field-initializers and -Wunused-parameter.
333 2. The tests for quantifiers being too big (greater than 65535) were being
334 applied after reading the number, and stupidly assuming that integer
335 overflow would give a negative number. The tests are now applied as the
336 numbers are read.
338 3. Tidy code in pcre_exec.c where two branches that used to be different are
339 now the same.
341 4. The JIT compiler did not generate match limit checks for certain
342 bracketed expressions with quantifiers. This may lead to exponential
343 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
344 issue should be resolved now.
346 5. Fixed an issue, which occures when nested alternatives are optimized
347 with table jumps.
349 6. Inserted two casts and changed some ints to size_t in the light of some
350 reported 64-bit compiler warnings (Bugzilla 1477).
352 7. Fixed a bug concerned with zero-minimum possessive groups that could match
353 an empty string, which sometimes were behaving incorrectly in the
354 interpreter (though correctly in the JIT matcher). This pcretest input is
355 an example:
357 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
360 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
361 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
362 for an empty string was breaking the inner loop and carrying on at a lower
363 level, when possessive repeated groups should always return to a higher
364 level as they have no backtrack points in them. The empty string test now
365 occurs at the outer level.
367 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
368 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
370 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
371 doesn't).
373 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
374 to hold a pattern to be incorrectly computed (too small) when there were
375 named back references to duplicated names. This could cause "internal
376 error: code overflow" or "double free or corruption" or other memory
377 handling errors.
379 11. When named subpatterns had the same prefixes, back references could be
380 confused. For example, in this pattern:
382 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
384 the reference to 'Name' was incorrectly treated as a reference to a
385 duplicate name.
387 12. A pattern such as /^s?c/mi8 where the optional character has more than
388 one "other case" was incorrectly compiled such that it would only try to
389 match starting at "c".
391 13. When a pattern starting with \s was studied, VT was not included in the
392 list of possible starting characters; this should have been part of the
393 8.34/18 patch.
395 14. If a character class started [\Qx]... where x is any character, the class
396 was incorrectly terminated at the ].
398 15. If a pattern that started with a caseless match for a character with more
399 than one "other case" was studied, PCRE did not set up the starting code
400 unit bit map for the list of possible characters. Now it does. This is an
401 optimization improvement, not a bug fix.
403 16. The Unicode data tables have been updated to Unicode 7.0.0.
405 17. Fixed a number of memory leaks in pcregrep.
407 18. Avoid a compiler warning (from some compilers) for a function call with
408 a cast that removes "const" from an lvalue by using an intermediate
409 variable (to which the compiler does not object).
411 19. Incorrect code was compiled if a group that contained an internal recursive
412 back reference was optional (had quantifier with a minimum of zero). This
413 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
414 caused segmentation faults because of stack overflows at compile time.
416 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
417 group that is quantified with an indefinite repeat, caused a compile-time
418 loop which used up all the system stack and provoked a segmentation fault.
419 This was not the same bug as 19 above.
421 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
422 Patch by Mike Frysinger.
425 Version 8.35 04-April-2014
426 --------------------------
428 1. A new flag is set, when property checks are present in an XCLASS.
429 When this flag is not set, PCRE can perform certain optimizations
430 such as studying these XCLASS-es.
432 2. The auto-possessification of character sets were improved: a normal
433 and an extended character set can be compared now. Furthermore
434 the JIT compiler optimizes more character set checks.
436 3. Got rid of some compiler warnings for potentially uninitialized variables
437 that show up only when compiled with -O2.
439 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
440 of a match later then the end of the match. The pcretest program was not
441 handling the case sensibly - it was outputting from the start to the next
442 binary zero. It now reports this situation in a message, and outputs the
443 text from the end to the start.
445 5. Fast forward search is improved in JIT. Instead of the first three
446 characters, any three characters with fixed position can be searched.
447 Search order: first, last, middle.
449 6. Improve character range checks in JIT. Characters are read by an inprecise
450 function now, which returns with an unknown value if the character code is
451 above a certain threshold (e.g: 256). The only limitation is that the value
452 must be bigger than the threshold as well. This function is useful when
453 the characters above the threshold are handled in the same way.
455 7. The macros whose names start with RAWUCHAR are placeholders for a future
456 mode in which only the bottom 21 bits of 32-bit data items are used. To
457 make this more memorable for those maintaining the code, the names have
458 been changed to start with UCHAR21, and an extensive comment has been added
459 to their definition.
461 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
462 to the export list in Makefile.am (they were accidentally omitted from the
463 8.34 tarball).
465 9. The informational output from pcretest used the phrase "starting byte set"
466 which is inappropriate for the 16-bit and 32-bit libraries. As the output
467 for "first char" and "need char" really means "non-UTF-char", I've changed
468 "byte" to "char", and slightly reworded the output. The documentation about
469 these values has also been (I hope) clarified.
471 10. Another JIT related optimization: use table jumps for selecting the correct
472 backtracking path, when more than four alternatives are present inside a
473 bracket.
475 11. Empty match is not possible, when the minimum length is greater than zero,
476 and there is no \K in the pattern. JIT should avoid empty match checks in
477 such cases.
479 12. In a caseless character class with UCP support, when a character with more
480 than one alternative case was not the first character of a range, not all
481 the alternative cases were added to the class. For example, s and \x{17f}
482 are both alternative cases for S: the class [RST] was handled correctly,
483 but [R-T] was not.
485 13. The configure.ac file always checked for pthread support when JIT was
486 enabled. This is not used in Windows, so I have put this test inside a
487 check for the presence of windows.h (which was already tested for).
489 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
490 The algorithm provides a way to skip certain starting offsets, and usually
491 faster than linear prefix searches.
493 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
494 as for 'fr_FR' and 'french'. For some reason, however, it then used the
495 Windows-specific input and output files, which have 'french' screwed in.
496 So this could never have worked. One of the problems with locales is that
497 they aren't always the same. I have now updated RunTest so that it checks
498 the output of the locale test (test 3) against three different output
499 files, and it allows the test to pass if any one of them matches. With luck
500 this should make the test pass on some versions of Solaris where it was
501 failing. Because of the uncertainty, the script did not used to stop if
502 test 3 failed; it now does. If further versions of a French locale ever
503 come to light, they can now easily be added.
505 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
506 there was a message during ./configure, but it did not stop. This now
507 provokes an error. The invalid example in README has been corrected.
508 If a value less than the minimum is given, the minimum value has always
509 been used, but now a warning is given.
511 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
512 was a bug in the test system, which is now fixed. Also, the list of various
513 configurations that are tested for each release did not have one with both
514 16/32 bits and --enable-bar-anycrlf. It now does.
516 18. pcretest was missing "-C bsr" for displaying the \R default setting.
518 19. Little endian PowerPC systems are supported now by the JIT compiler.
520 20. The fast forward newline mechanism could enter to an infinite loop on
521 certain invalid UTF-8 input. Although we don't support these cases
522 this issue can be fixed by a performance optimization.
524 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
525 not take account if existing stack usage. There is now a new global
526 variable called pcre_stack_guard that can be set to point to an external
527 function to check stack availability. It is called at the start of
528 processing every parenthesized group.
530 22. A typo in the code meant that in ungreedy mode the max/min qualifier
531 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
532 not match "ab".
534 23. When UTF was disabled, the JIT program reported some incorrect compile
535 errors. These messages are silenced now.
537 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
538 compiler.
540 25. Change all the temporary files used in RunGrepTest to be different to those
541 used by RunTest so that the tests can be run simultaneously, for example by
542 "make -j check".
545 Version 8.34 15-December-2013
546 -----------------------------
548 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
549 executable memory. Patch inspired by Carsten Klein.
551 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
552 this macro is never tested and has no effect, because the work to support
553 coverage involves only compiling and linking options and special targets in
554 the Makefile. The comment in config.h implied that defining the macro would
555 enable coverage support, which is totally false. There was also support for
556 setting this macro in the CMake files (my fault, I just copied it from
557 configure). SUPPORT_GCOV has now been removed.
559 3. Make a small performance improvement in strlen16() and strlen32() in
560 pcretest.
562 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
563 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
564 these cases). There was also one in pcretest.c.
566 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
568 6. In UTF mode, the code for checking whether a group could match an empty
569 string (which is used for indefinitely repeated groups to allow for
570 breaking an infinite loop) was broken when the group contained a repeated
571 negated single-character class with a character that occupied more than one
572 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
573 UTF-8 mode). The effect was undefined: the group might or might not be
574 deemed as matching an empty string, or the program might have crashed.
576 7. The code for checking whether a group could match an empty string was not
577 recognizing that \h, \H, \v, \V, and \R must match a character.
579 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
580 an empty string. If it can, pcretest shows this in its information output.
582 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
583 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
584 matched by pcre_exec() without using JIT:
586 (a) If the rest of the pattern did not match after a maximal run of
587 grapheme clusters, the code for backing up to try with fewer of them
588 did not always back up over a full grapheme when characters that do not
589 have the modifier quality were involved, e.g. Hangul syllables.
591 (b) If the match point in a subject started with modifier character, and
592 there was no match, the code could incorrectly back up beyond the match
593 point, and potentially beyond the first character in the subject,
594 leading to a segfault or an incorrect match result.
596 10. A conditional group with an assertion condition could lead to PCRE
597 recording an incorrect first data item for a match if no other first data
598 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
599 first data item, and therefore matched "ca" after "c" instead of at the
600 start.
602 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
603 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
605 12. The source of pcregrep now includes z/OS-specific code so that it can be
606 compiled for z/OS as part of the special z/OS distribution.
608 13. Added the -T and -TM options to pcretest.
610 14. The code in pcre_compile.c for creating the table of named capturing groups
611 has been refactored. Instead of creating the table dynamically during the
612 actual compiling pass, the information is remembered during the pre-compile
613 pass (on the stack unless there are more than 20 named groups, in which
614 case malloc() is used) and the whole table is created before the actual
615 compile happens. This has simplified the code (it is now nearly 150 lines
616 shorter) and prepared the way for better handling of references to groups
617 with duplicate names.
619 15. A back reference to a named subpattern when there is more than one of the
620 same name now checks them in the order in which they appear in the pattern.
621 The first one that is set is used for the reference. Previously only the
622 first one was inspected. This change makes PCRE more compatible with Perl.
624 16. Unicode character properties were updated from Unicode 6.3.0.
626 17. The compile-time code for auto-possessification has been refactored, based
627 on a patch by Zoltan Herczeg. It now happens after instead of during
628 compilation. The code is cleaner, and more cases are handled. The option
629 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
630 options in pcretest are provided to set it. It can also be set by
631 (*NO_AUTO_POSSESS) at the start of a pattern.
633 18. The character VT has been added to the default ("C" locale) set of
634 characters that match \s and are generally treated as white space,
635 following this same change in Perl 5.18. There is now no difference between
636 "Perl space" and "POSIX space". Whether VT is treated as white space in
637 other locales depends on the locale.
639 19. The code for checking named groups as conditions, either for being set or
640 for being recursed, has been refactored (this is related to 14 and 15
641 above). Processing unduplicated named groups should now be as fast at
642 numerical groups, and processing duplicated groups should be faster than
643 before.
645 20. Two patches to the CMake build system, by Alexander Barkov:
647 (1) Replace the "source" command by "." in CMakeLists.txt because
648 "source" is a bash-ism.
650 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
651 without these the CMake build does not work on Solaris.
653 21. Perl has changed its handling of \8 and \9. If there is no previously
654 encountered capturing group of those numbers, they are treated as the
655 literal characters 8 and 9 instead of a binary zero followed by the
656 literals. PCRE now does the same.
658 22. Following Perl, added \o{} to specify codepoints in octal, making it
659 possible to specify values greater than 0777 and also making them
660 unambiguous.
662 23. Perl now gives an error for missing closing braces after \x{... instead of
663 treating the string as literal. PCRE now does the same.
665 24. RunTest used to grumble if an inappropriate test was selected explicitly,
666 but just skip it when running all tests. This make it awkward to run ranges
667 of tests when one of them was inappropriate. Now it just skips any
668 inappropriate tests, as it always did when running all tests.
670 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
671 character types such as \d or \w, too many callouts were inserted, and the
672 data that they returned was rubbish.
674 26. In UCP mode, \s was not matching two of the characters that Perl matches,
675 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
676 were matched by \h. The code has now been refactored so that the lists of
677 the horizontal and vertical whitespace characters used for \h and \v (which
678 are defined only in one place) are now also used for \s.
680 27. Add JIT support for the 64 bit TileGX architecture.
681 Patch by Jiong Wang (Tilera Corporation).
683 28. Possessive quantifiers for classes (both explicit and automatically
684 generated) now use special opcodes instead of wrapping in ONCE brackets.
686 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
687 (because it's meaningless), this was not happening when PCRE_CASELESS was
688 set. Not wrong, but inefficient.
690 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
691 properties for \w, \d, etc) is present in a test regex. Otherwise if the
692 test contains no characters greater than 255, Perl doesn't realise it
693 should be using Unicode semantics.
695 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
696 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
697 does in Unicode mode.
699 32. Added the "forbid" facility to pcretest so that putting tests into the
700 wrong test files can sometimes be quickly detected.
702 33. There is now a limit (default 250) on the depth of nesting of parentheses.
703 This limit is imposed to control the amount of system stack used at compile
704 time. It can be changed at build time by --with-parens-nest-limit=xxx or
705 the equivalent in CMake.
707 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
708 errors. Perl warns for these when in warning mode, but PCRE has no facility
709 for giving warnings.
711 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
712 However, this was not working for (?!) because it is optimized to (*FAIL),
713 for which PCRE does not allow quantifiers. The optimization is now disabled
714 when a quantifier follows (?!). I can't see any use for this, but it makes
715 things uniform.
717 36. Perl no longer allows group names to start with digits, so I have made this
718 change also in PCRE. It simplifies the code a bit.
720 37. In extended mode, Perl ignores spaces before a + that indicates a
721 possessive quantifier. PCRE allowed a space before the quantifier, but not
722 before the possessive +. It now does.
724 38. The use of \K (reset reported match start) within a repeated possessive
725 group such as (a\Kb)*+ was not working.
727 40. Document that the same character tables must be used at compile time and
728 run time, and that the facility to pass tables to pcre_exec() and
729 pcre_dfa_exec() is for use only with saved/restored patterns.
731 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
732 features for Builds with MSVC:
734 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
735 builds). The .rc files can be used to set FileDescription and many other
736 attributes.
738 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
739 This allows higher-level build scripts which want .pdb files to avoid
740 hard-coding the exact files needed."
742 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
743 mean "start of word" and "end of word", respectively, as a transition aid.
745 43. A minimizing repeat of a class containing codepoints greater than 255 in
746 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
747 compiled to use the heap for recursion.
749 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
750 is configured.
753 Version 8.33 28-May-2013
754 ------------------------
756 1. Added 'U' to some constants that are compared to unsigned integers, to
757 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
758 variables that are added to signed variables, to ensure the result is
759 signed and can be negated.
761 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
762 CMake config files.
764 3. Revise the creation of config.h.generic so that all boolean macros are
765 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
766 overriding via -D on the command line possible.
768 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
769 to unsigned int is reported to make a quite noticeable speed difference in
770 a specific Windows environment. Testing on Linux did also appear to show
771 some benefit (and it is clearly not harmful). Also fixed the definition of
772 Xop which should be unsigned.
774 5. Related to (4), changing the definition of the intermediate variable cc
775 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
776 improvements.
778 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
779 unnecessary spaces.
781 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
782 and later.
784 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
786 9. Optimizing fast_forward_start_bits in JIT.
788 10. Adding support for callouts in JIT, and fixing some issues revealed
789 during this work. Namely:
791 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
793 (b) Minimum length was not checked before the matching is started.
795 11. The value of capture_last that is passed to callouts was incorrect in some
796 cases when there was a capture on one path that was subsequently abandoned
797 after a backtrack. Also, the capture_last value is now reset after a
798 recursion, since all captures are also reset in this case.
800 12. The interpreter no longer returns the "too many substrings" error in the
801 case when an overflowing capture is in a branch that is subsequently
802 abandoned after a backtrack.
804 13. In the pathological case when an offset vector of size 2 is used, pcretest
805 now prints out the matched string after a yield of 0 or 1.
807 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
808 Only supported by the JIT compiler at the moment.
810 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
812 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
813 offset of the starting point of the matching process, provided the offsets
814 vector is large enough.
816 17. The \A escape now records a lookbehind value of 1, though its execution
817 does not actually inspect the previous character. This is to ensure that,
818 in partial multi-segment matching, at least one character from the old
819 segment is retained when a new segment is processed. Otherwise, if there
820 are no lookbehinds in the pattern, \A might match incorrectly at the start
821 of a new segment.
823 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
825 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
826 gives some modest performance improvement in 8-bit mode.
828 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
829 be expressed in certain programming languages using Universal Character
830 Names.
832 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
833 which points out that "non characters" are not "characters that may not
834 appear in Unicode strings" but rather "characters that are reserved for
835 internal use and have only local meaning".
837 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
838 there was a conditional group that depended on an assertion, if the
839 assertion was false, the callout that immediately followed the alternation
840 in the condition was skipped when pcre_exec() was used for matching.
842 23. Allow an explicit callout to be inserted before an assertion that is the
843 condition for a conditional group, for compatibility with automatic
844 callouts, which always insert a callout at this point.
846 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
847 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
849 25. (*PRUNE) is now supported by the JIT compiler.
851 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
853 27. Fix the case where there are two or more SKIPs with arguments that may be
854 ignored.
856 28. (*SKIP) is now supported by the JIT compiler.
858 29. (*THEN) is now supported by the JIT compiler.
860 30. Update RunTest with additional test selector options.
862 31. The way PCRE handles backtracking verbs has been changed in two ways.
864 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
865 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
866 backtracking. In some cases this makes it more Perl-compatible, but Perl's
867 rather obscure rules do not always do the same thing.
869 (2) Previously, backtracking verbs were confined within assertions. This is
870 no longer the case for positive assertions, except for (*ACCEPT). Again,
871 this sometimes improves Perl compatibility, and sometimes does not.
873 32. A number of tests that were in test 2 because Perl did things differently
874 have been moved to test 1, because either Perl or PCRE has changed, and
875 these tests are now compatible.
877 32. Backtracking control verbs are now handled in the same way in JIT and
878 interpreter.
880 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
881 contained a forward subroutine reference caused a compile error.
883 34. Auto-detect and optimize limited repetitions in JIT.
885 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
886 blocking (*UTF) etc.
888 36. In the interpreter, maximizing pattern repetitions for characters and
889 character types now use tail recursion, which reduces stack usage.
891 37. The value of the max lookbehind was not correctly preserved if a compiled
892 and saved regex was reloaded on a host of different endianness.
894 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
895 of the compiled pattern block, expand the flags field from 16 to 32 bits
896 because it was almost full.
898 39. Try madvise first before posix_madvise.
900 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
901 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
902 original change locked out all matches of empty strings. This has been
903 changed so that one match of an empty string per line is recognized.
904 Subsequent searches on the same line (for colouring or for --only-matching,
905 for example) do not recognize empty strings.
907 41. Applied a user patch to fix a number of spelling mistakes in comments.
909 42. Data lines longer than 65536 caused pcretest to crash.
911 43. Clarified the data type for length and startoffset arguments for pcre_exec
912 and pcre_dfa_exec in the function-specific man pages, where they were
913 explicitly stated to be in bytes, never having been updated. I also added
914 some clarification to the pcreapi man page.
916 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
917 a segmentation fault.
920 Version 8.32 30-November-2012
921 -----------------------------
923 1. Improved JIT compiler optimizations for first character search and single
924 character iterators.
926 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
927 Patch by Daniel Richard G.
929 3. Single character iterator optimizations in the JIT compiler.
931 4. Improved JIT compiler optimizations for character ranges.
933 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
934 Reported by Giuseppe D'Angelo.
936 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
937 of a line, was being set incorrectly in cases where .* appeared inside
938 atomic brackets at the start of a pattern, or where there was a subsequent
939 *PRUNE or *SKIP.
941 7. Improved instruction cache flush for POWER/PowerPC.
942 Patch by Daniel Richard G.
944 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
945 grep:
947 (a) There is now no limit to the number of patterns to be matched.
949 (b) An error is given if a pattern is too long.
951 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
952 are now supported.
954 (d) --exclude-from and --include-from (multiple use) have been added.
956 (e) Exclusions and inclusions now apply to all files and directories, not
957 just to those obtained from scanning a directory recursively.
959 (f) Multiple uses of -f and --file-list are now supported.
961 (g) In a Windows environment, the default for -d has been changed from
962 "read" (the GNU grep default) to "skip", because otherwise the presence
963 of a directory in the file list provokes an error.
965 (h) The documentation has been revised and clarified in places.
967 9. Improve the matching speed of capturing brackets.
969 10. Changed the meaning of \X so that it now matches a Unicode extended
970 grapheme cluster.
972 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
973 out POSIX threads when JIT support is configured.
975 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
977 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
978 structure could go wrong in environments where size_t is not the same size
979 as int.
981 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
982 set.
984 15. The EBCDIC support had decayed; later updates to the code had included
985 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
986 general tidy up of EBCDIC-related issues, and the documentation was also
987 not quite right. There is now a test that can be run on ASCII systems to
988 check some of the EBCDIC-related things (but is it not a full test).
990 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
991 in a small tidy to the code.
993 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
995 18. If the --only-matching (-o) option in pcregrep is specified multiple
996 times, each one causes appropriate output. For example, -o1 -o2 outputs the
997 substrings matched by the 1st and 2nd capturing parentheses. A separating
998 string can be specified by --om-separator (default empty).
1000 19. Improving the first n character searches.
1002 20. Turn case lists for horizontal and vertical white space into macros so that
1003 they are defined only once.
1005 21. This set of changes together give more compatible Unicode case-folding
1006 behaviour for characters that have more than one other case when UCP
1007 support is available.
1009 (a) The Unicode property table now has offsets into a new table of sets of
1010 three or more characters that are case-equivalent. The MultiStage2.py
1011 script that generates these tables (the pcre_ucd.c file) now scans
1012 CaseFolding.txt instead of UnicodeData.txt for character case
1013 information.
1015 (b) The code for adding characters or ranges of characters to a character
1016 class has been abstracted into a generalized function that also handles
1017 case-independence. In UTF-mode with UCP support, this uses the new data
1018 to handle characters with more than one other case.
1020 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1021 whose other case is greater than 256 are now correctly matched
1022 caselessly. Previously, the high codepoint matched the low one, but not
1023 vice versa.
1025 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1026 of the new class addition function, using character lists defined as
1027 macros alongside the case definitions of 20 above.
1029 (e) Caseless back references now work with characters that have more than
1030 one other case.
1032 (f) General caseless matching of characters with more than one other case
1033 is supported.
1035 22. Unicode character properties were updated from Unicode 6.2.0
1037 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1039 24. Add support for 32-bit character strings, and UTF-32
1041 25. Major JIT compiler update (code refactoring and bugfixing).
1042 Experimental Sparc 32 support is added.
1044 26. Applied a modified version of Daniel Richard G's patch to create
1045 pcre.h.generic and config.h.generic by "make" instead of in the
1046 PrepareRelease script.
1048 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1049 pcre_compile.c when checking for a zero character.
1051 28. Introducing a native interface for JIT. Through this interface, the compiled
1052 machine code can be directly executed. The purpose of this interface is to
1053 provide fast pattern matching, so several sanity checks are not performed.
1054 However, feature tests are still performed. The new interface provides
1055 1.4x speedup compared to the old one.
1057 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1058 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1059 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1061 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1062 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1063 the "old" RFC 2279). Instead, it ended up passing a negative length to
1064 pcre_exec().
1066 31. Add support for GCC's visibility feature to hide internal functions.
1068 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1069 "unknown -C option" after outputting 0 or 1.
1071 33. There is now support for generating a code coverage report for the test
1072 suite in environments where gcc is the compiler and lcov is installed. This
1073 is mainly for the benefit of the developers.
1075 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1076 unaddressable using valgrind annotations, allowing valgrind to detect
1077 invalid memory accesses. This is mainly for the benefit of the developers.
1079 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1081 26. Give configure error if --enable-cpp but no C++ compiler found.
1084 Version 8.31 06-July-2012
1085 -------------------------
1087 1. Fixing a wrong JIT test case and some compiler warnings.
1089 2. Removed a bashism from the RunTest script.
1091 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1092 to unsigned type, result still unsigned" that was given by an MS compiler
1093 on encountering the code "-sizeof(xxx)".
1095 4. Partial matching support is added to the JIT compiler.
1097 5. Fixed several bugs concerned with partial matching of items that consist
1098 of more than one character:
1100 (a) /^(..)\1/ did not partially match "aba" because checking references was
1101 done on an "all or nothing" basis. This also applied to repeated
1102 references.
1104 (b) \R did not give a hard partial match if \r was found at the end of the
1105 subject.
1107 (c) \X did not give a hard partial match after matching one or more
1108 characters at the end of the subject.
1110 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1111 a partial match for the string "\r".
1113 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1114 a partial match for a CR character at the end of the subject string.
1116 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1117 running pcretest, the text "(JIT)" added to the output whenever JIT is
1118 actually used to run the match.
1120 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1121 or /S+[+] with a digit between 1 and 7.
1123 8. OP_NOT now supports any UTF character not just single-byte ones.
1125 9. (*MARK) control verb is now supported by the JIT compiler.
1127 10. The command "./RunTest list" lists the available tests without actually
1128 running any of them. (Because I keep forgetting what they all are.)
1132 12. Applied a (slightly modified) user-supplied patch that improves performance
1133 when the heap is used for recursion (compiled with --disable-stack-for-
1134 recursion). Instead of malloc and free for each heap frame each time a
1135 logical recursion happens, frames are retained on a chain and re-used where
1136 possible. This sometimes gives as much as 30% improvement.
1138 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1139 call.
1141 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1143 15. It is now possible to link pcretest with libedit as an alternative to
1144 libreadline.
1146 16. (*COMMIT) control verb is now supported by the JIT compiler.
1148 17. The Unicode data tables have been updated to Unicode 6.1.0.
1150 18. Added --file-list option to pcregrep.
1152 19. Added binary file support to pcregrep, including the -a, --binary-files,
1153 -I, and --text options.
1155 20. The madvise function is renamed for posix_madvise for QNX compatibility
1156 reasons. Fixed by Giuseppe D'Angelo.
1158 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1160 22. Changed the option for creating symbolic links for 16-bit man pages from
1161 -s to -sf so that re-installing does not cause issues.
1163 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1165 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1166 matches in certain environments (the workspace was not being correctly
1167 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1168 some of the workspace data at the beginning of a restart.
1170 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1171 was not doing so when it should - probably a typo introduced by SVN 528
1172 (change 8.10/14).
1174 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1175 \w+ when the character tables indicated that \x{c4} was a word character.
1176 There were several related cases, all because the tests for doing a table
1177 lookup were testing for characters less than 127 instead of 255.
1179 27. If a pattern contains capturing parentheses that are not used in a match,
1180 their slots in the ovector are set to -1. For those that are higher than
1181 any matched groups, this happens at the end of processing. In the case when
1182 there were back references that the ovector was too small to contain
1183 (causing temporary malloc'd memory to be used during matching), and the
1184 highest capturing number was not used, memory off the end of the ovector
1185 was incorrectly being set to -1. (It was using the size of the temporary
1186 memory instead of the true size.)
1188 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1189 ovector size, it uses memory at the end of the block that it has got.
1191 29. Check for an overlong MARK name and give an error at compile time. The
1192 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1194 30. JIT compiler update.
1196 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1197 Rill for the patch.
1199 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1201 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1203 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1204 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1205 was enabled.
1207 35. Improve JIT code generation for greedy plus quantifier.
1209 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1210 1 to "aa" instead of to an empty string. The bug affected repeated groups
1211 that could potentially match an empty string.
1213 37. Optimizing single character iterators in JIT.
1215 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1216 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1217 codepoints that are too big for the mode are faulted, and in a UTF mode,
1218 disallowed codepoints are also faulted.
1220 39. If PCRE was compiled with UTF support, in three places in the DFA
1221 matcher there was code that should only have been obeyed in UTF mode, but
1222 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1223 processing when bytes with values greater than 127 were present. In 16-bit
1224 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1225 both cases the values are those that cannot be the first data item in a UTF
1226 character. The three items that might have provoked this were recursions,
1227 possessively repeated groups, and atomic groups.
1229 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1230 and pcregrep, because some OS require shared objects to be explicitly
1231 passed to ld, causing the link step to fail if they are not.
1233 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1234 mode, patterns that started with \h* or \R* might be incorrectly matched.
1237 Version 8.30 04-February-2012
1238 -----------------------------
1240 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1241 name is defined in ctype.h.
1243 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1244 only in quite long subpatterns.
1246 3. Removed the function pcre_info(), which has been obsolete and deprecated
1247 since it was replaced by pcre_fullinfo() in February 2000.
1249 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1250 match a (*MARK), and the match failed at the start of the subject, a
1251 reference to memory before the start of the subject could occur. This bug
1252 was introduced by fix 17 of release 8.21.
1254 5. A reference to an unset group with zero minimum repetition was giving
1255 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1256 /(another)?(\1?)test/ matched against "hello world test". This bug was
1257 introduced in release 8.13.
1259 6. Add support for 16-bit character strings (a large amount of work involving
1260 many changes and refactorings).
1262 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1263 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1264 from a file.
1266 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1267 rounding is not applied in this particular case).
1269 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1270 if they appear, or are escaped, in patterns.
1272 10. Get rid of a number of -Wunused-but-set-variable warnings.
1274 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1275 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1276 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1277 also returns the mark "x". This bug applied to capturing parentheses,
1278 non-capturing parentheses, and atomic parentheses. It also applied to some
1279 assertions.
1281 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1282 information out of configure.ac instead of relying on pcre.h.generic, which
1283 is not stored in the repository.
1285 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1286 -lreadline.
1288 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1290 15. Applied Graycode's patch to put the top-level frame on the stack rather
1291 than the heap when not using the stack for recursion. This gives a
1292 performance improvement in many cases when recursion is not deep.
1294 16. Experimental code added to "pcretest -C" to output the stack frame size.
1297 Version 8.21 12-Dec-2011
1298 ------------------------
1300 1. Updating the JIT compiler.
1302 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1303 are added as well.
1305 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1306 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1307 calling _pcre_jit_exec. Some extra comments are added.
1309 4. (*MARK) settings inside atomic groups that do not contain any capturing
1310 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1311 was introduced by change 18 for 8.20.
1313 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1314 ECMA-262 standard.
1316 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1317 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1318 This bug was probably introduced by change 9 of 8.13.
1320 7. While fixing 6 above, I noticed that a number of other items were being
1321 incorrectly rejected as "not fixed length". This arose partly because newer
1322 opcodes had not been added to the fixed-length checking code. I have (a)
1323 corrected the bug and added tests for these items, and (b) arranged for an
1324 error to occur if an unknown opcode is encountered while checking for fixed
1325 length instead of just assuming "not fixed length". The items that were
1326 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1327 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1328 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1330 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1331 being incorrectly compiled and would have given unpredicatble results.
1333 9. A possessively repeated subpattern with minimum repeat count greater than
1334 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1335 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1336 the first (A) could occur when it should not.
1338 10. Add a cast and remove a redundant test from the code.
1340 11. JIT should use pcre_malloc/pcre_free for allocation.
1342 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1343 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1344 is anything other than /usr, -L is still shown).
1346 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1348 14. Perl does not support \N without a following name in a [] class; PCRE now
1349 also gives an error.
1351 15. If a forward reference was repeated with an upper limit of around 2000,
1352 it caused the error "internal error: overran compiling workspace". The
1353 maximum number of forward references (including repeats) was limited by the
1354 internal workspace, and dependent on the LINK_SIZE. The code has been
1355 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1356 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1357 of around 200,000 forward references. While doing this, I also speeded up
1358 the filling in of repeated forward references.
1360 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1361 incorrectly expecting the subject to contain another "a" after the start.
1363 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1364 in the match, the SKIP should be ignored. This was not happening; instead
1365 the SKIP was being treated as NOMATCH. For patterns such as
1366 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1367 tested.
1369 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1370 now much more compatible with Perl, in particular in cases where the result
1371 is a non-match for a non-anchored pattern. For example, if
1372 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1373 "m", where previously it did not return a name. A side effect of this
1374 change is that for partial matches, the last encountered mark name is
1375 returned, as for non matches. A number of tests that were previously not
1376 Perl-compatible have been moved into the Perl-compatible test files. The
1377 refactoring has had the pleasing side effect of removing one argument from
1378 the match() function, thus reducing its stack requirements.
1380 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1381 subsequent uses of /S (without +) incorrectly behaved like /S+.
1383 21. Retrieve executable code size support for the JIT compiler and fixing
1384 some warnings.
1386 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1387 not work when the shorter character appeared right at the end of the
1388 subject string.
1390 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1391 systems.
1393 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1394 output it when the /M option is used in pcretest.
1396 25. The CheckMan script was not being included in the distribution. Also, added
1397 an explicit "perl" to run Perl scripts from the PrepareRelease script
1398 because this is reportedly needed in Windows.
1400 26. If study data was being save in a file and studying had not found a set of
1401 "starts with" bytes for the pattern, the data written to the file (though
1402 never used) was taken from uninitialized memory and so caused valgrind to
1403 complain.
1405 27. Updated RunTest.bat as provided by Sheri Pierce.
1407 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1409 29. Computation of memory usage for the table of capturing group names was
1410 giving an unnecessarily large value.
1413 Version 8.20 21-Oct-2011
1414 ------------------------
1416 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1417 a POSIX class. After further experiments with Perl, which convinced me that
1418 Perl has bugs and confusions, a closing square bracket is no longer allowed
1419 in a POSIX name. This bug also affected patterns with classes that started
1420 with full stops.
1422 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1423 captured substring, but while checking the failing first alternative,
1424 substring 1 is temporarily captured. If the output vector supplied to
1425 pcre_exec() was not big enough for this capture, the yield of the function
1426 was still zero ("insufficient space for captured substrings"). This cannot
1427 be totally fixed without adding another stack variable, which seems a lot
1428 of expense for a edge case. However, I have improved the situation in cases
1429 such as /(a)(b)x|abc/ matched against "abc", where the return code
1430 indicates that fewer than the maximum number of slots in the ovector have
1431 been set.
1433 3. Related to (2) above: when there are more back references in a pattern than
1434 slots in the output vector, pcre_exec() uses temporary memory during
1435 matching, and copies in the captures as far as possible afterwards. It was
1436 using the entire output vector, but this conflicts with the specification
1437 that only 2/3 is used for passing back captured substrings. Now it uses
1438 only the first 2/3, for compatibility. This is, of course, another edge
1439 case.
1441 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1442 main code base, and can be used by building with --enable-jit. When this is
1443 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1444 runtime --no-jit option is given.
1446 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1447 ovector, the return from the function was zero, implying that there were
1448 other matches that did not fit. The correct "exactly full" value is now
1449 returned.
1451 6. If a subpattern that was called recursively or as a subroutine contained
1452 (*PRUNE) or any other control that caused it to give a non-standard return,
1453 invalid errors such as "Error -26 (nested recursion at the same subject
1454 position)" or even infinite loops could occur.
1456 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1457 computing the minimum length on reaching *ACCEPT, and so ended up with the
1458 wrong value of 1 rather than 0. Further investigation indicates that
1459 computing a minimum subject length in the presence of *ACCEPT is difficult
1460 (think back references, subroutine calls), and so I have changed the code
1461 so that no minimum is registered for a pattern that contains *ACCEPT.
1463 8. If (*THEN) was present in the first (true) branch of a conditional group,
1464 it was not handled as intended. [But see 16 below.]
1466 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1467 Sheri Pierce.
1469 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1470 the first byte in a match must be "a".
1472 11. Change 17 for 8.13 increased the recursion depth for patterns like
1473 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1474 pattern contains any instances of (*THEN). If it does not, the old
1475 optimizations are restored. It would be nice to do this on a per-group
1476 basis, but at the moment that is not feasible.
1478 12. In some environments, the output of pcretest -C is CRLF terminated. This
1479 broke RunTest's code that checks for the link size. A single white space
1480 character after the value is now allowed for.
1482 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1483 For "fr", it uses the Windows-specific input and output files.
1485 14. If (*THEN) appeared in a group that was called recursively or as a
1486 subroutine, it did not work as intended. [But see next item.]
1488 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1489 pattern fragments (but not containing any | characters). If A and B are
1490 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1491 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1492 D. In other words, Perl considers parentheses that do not contain any |
1493 characters to be part of a surrounding alternative, whereas PCRE was
1494 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1495 differently. PCRE now behaves in the same way as Perl, except in the case
1496 of subroutine/recursion calls such as (?1) which have in any case always
1497 been different (but PCRE had them first :-).
1499 16. Related to 15 above: Perl does not treat the | in a conditional group as
1500 creating alternatives. Such a group is treated in the same way as an
1501 ordinary group without any | characters when processing (*THEN). PCRE has
1502 been changed to match Perl's behaviour.
1504 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1505 RunGrepTest script failed.
1507 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1508 inevitable for groups that contain captures, but it can lead to a lot of
1509 stack use in large patterns. The old behaviour has been restored for atomic
1510 groups that do not contain any capturing parentheses.
1512 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1513 suppress the check for a minimum subject length at run time. (If it was
1514 given to pcre_exec() or pcre_dfa_exec() it did work.)
1516 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1517 fail to work when decoding hex characters in data strings in EBCDIC
1518 environments.
1520 21. It appears that in at least one Mac OS environment, the isxdigit() function
1521 is implemented as a macro that evaluates to its argument more than once,
1522 contravening the C 90 Standard (I haven't checked a later standard). There
1523 was an instance in pcretest which caused it to go wrong when processing
1524 \x{...} escapes in subject strings. The has been rewritten to avoid using
1525 things like p++ in the argument of isxdigit().
1528 Version 8.13 16-Aug-2011
1529 ------------------------
1531 1. The Unicode data tables have been updated to Unicode 6.0.0.
1533 2. Two minor typos in pcre_internal.h have been fixed.
1535 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1536 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1537 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1539 4. There were a number of related bugs in the code for matching backrefences
1540 caselessly in UTF-8 mode when codes for the characters concerned were
1541 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1542 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1543 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1544 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1545 2-byte code at the end of the subject (it thought there wasn't enough data
1546 left).
1548 5. Comprehensive information about what went wrong is now returned by
1549 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1550 as the output vector has at least 2 elements. The offset of the start of
1551 the failing character and a reason code are placed in the vector.
1553 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1554 now returned is for the first byte of the failing character, instead of the
1555 last byte inspected. This is an incompatible change, but I hope it is small
1556 enough not to be a problem. It makes the returned offset consistent with
1557 pcre_exec() and pcre_dfa_exec().
1559 7. pcretest now gives a text phrase as well as the error number when
1560 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1561 failure, the offset and reason code are output.
1563 8. When \R was used with a maximizing quantifier it failed to skip backwards
1564 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1565 back over a single character (\n). This seems wrong (because it treated the
1566 two characters as a single entity when going forwards), conflicts with the
1567 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1568 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1569 has been changed.
1571 9. Some internal refactoring has changed the processing so that the handling
1572 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1573 time (the PCRE_DOTALL option was changed this way some time ago: version
1574 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1575 which was always a bit of a fudge. It also means that there is one less
1576 argument for the match() function, which reduces its stack requirements
1577 slightly. This change also fixes an incompatibility with Perl: the pattern
1578 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1580 10. More internal refactoring has drastically reduced the number of recursive
1581 calls to match() for possessively repeated groups such as (abc)++ when
1582 using pcre_exec().
1584 11. While implementing 10, a number of bugs in the handling of groups were
1585 discovered and fixed:
1587 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1588 (a|)*(?1) gave a compile-time internal error.
1589 ((a|)+)+ did not notice that the outer group could match an empty string.
1590 (^a|^)+ was not marked as anchored.
1591 (.*a|.*)+ was not marked as matching at start or after a newline.
1593 12. Yet more internal refactoring has removed another argument from the match()
1594 function. Special calls to this function are now indicated by setting a
1595 value in a variable in the "match data" data block.
1597 13. Be more explicit in pcre_study() instead of relying on "default" for
1598 opcodes that mean there is no starting character; this means that when new
1599 ones are added and accidentally left out of pcre_study(), testing should
1600 pick them up.
1602 14. The -s option of pcretest has been documented for ages as being an old
1603 synonym of -m (show memory usage). I have changed it to mean "force study
1604 for every regex", that is, assume /S for every regex. This is similar to -i
1605 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1606 using it. It makes it easier to run collections of tests with and without
1607 study enabled, and thereby test pcre_study() more easily. All the standard
1608 tests are now run with and without -s (but some patterns can be marked as
1609 "never study" - see 20 below).
1611 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1612 restoration of the capturing data to the outer values was not happening
1613 correctly.
1615 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1616 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1617 pattern had matched an empty string, and so incorrectly returned a no
1618 match.
1620 17. There was optimizing code for the last branch of non-capturing parentheses,
1621 and also for the obeyed branch of a conditional subexpression, which used
1622 tail recursion to cut down on stack usage. Unfortunately, now that there is
1623 the possibility of (*THEN) occurring in these branches, tail recursion is
1624 no longer possible because the return has to be checked for (*THEN). These
1625 two optimizations have therefore been removed. [But see 8.20/11 above.]
1627 18. If a pattern containing \R was studied, it was assumed that \R always
1628 matched two bytes, thus causing the minimum subject length to be
1629 incorrectly computed because \R can also match just one byte.
1631 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1632 was incorrectly computed.
1634 20. If /S is present twice on a test pattern in pcretest input, it now
1635 *disables* studying, thereby overriding the use of -s on the command line
1636 (see 14 above). This is necessary for one or two tests to keep the output
1637 identical in both cases.
1639 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1640 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1642 22. When an atomic group that contained a capturing parenthesis was
1643 successfully matched, but the branch in which it appeared failed, the
1644 capturing was not being forgotten if a higher numbered group was later
1645 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1646 group 1 to "a", when in fact it should be unset. This applied to multi-
1647 branched capturing and non-capturing groups, repeated or not, and also to
1648 positive assertions (capturing in negative assertions does not happen
1649 in PCRE) and also to nested atomic groups.
1651 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1652 subject after a captured substring, to make it easier to tell which of a
1653 number of identical substrings has been captured.
1655 24. The way atomic groups are processed by pcre_exec() has been changed so that
1656 if they are repeated, backtracking one repetition now resets captured
1657 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1658 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1659 "aaa". Previously, it would have been "a". As part of this code
1660 refactoring, the way recursive calls are handled has also been changed.
1662 25. If an assertion condition captured any substrings, they were not passed
1663 back unless some other capturing happened later. For example, if
1664 (?(?=(a))a) was matched against "a", no capturing was returned.
1666 26. When studying a pattern that contained subroutine calls or assertions,
1667 the code for finding the minimum length of a possible match was handling
1668 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1669 group 1 called group 2 while simultaneously a separate group 2 called group
1670 1). A stack overflow occurred in this case. I have fixed this by limiting
1671 the recursion depth to 10.
1673 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1674 Fortmann. This supports explicit test numbers on the command line, and has
1675 argument validation and error reporting.
1677 28. An instance of \X with an unlimited repeat could fail if at any point the
1678 first character it looked at was a mark character.
1680 29. Some minor code refactoring concerning Unicode properties and scripts
1681 should reduce the stack requirement of match() slightly.
1683 30. Added the '=' option to pcretest to check the setting of unused capturing
1684 slots at the end of the pattern, which are documented as being -1, but are
1685 not included in the return count.
1687 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1688 compiled something random. Now it gives a compile-time error (as does
1689 Perl).
1691 32. A *MARK encountered during the processing of a positive assertion is now
1692 recorded and passed back (compatible with Perl).
1694 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1695 had alternative anchored branches, the search for a second match in a line
1696 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1697 matched the line "0102" twice. The same bug affected patterns that started
1698 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1699 twice.
1701 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1702 does, and because of capturing effects, quantifying parenthesized
1703 assertions may at times be useful. Quantifiers are now allowed for
1704 parenthesized assertions.
1706 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1708 36. \g was being checked for fancy things in a character class, when it should
1709 just be a literal "g".
1711 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1712 appearance of a nested POSIX class supersedes an apparent external class.
1713 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1714 unescaped square brackets may also appear as part of class names. For
1715 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1716 more like Perl. (But see 8.20/1 above.)
1718 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1719 was because it thought it was \N{name}, which is not supported).
1721 39. Add minix to OS list not supporting the -S option in pcretest.
1723 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1724 cannot analyze patterns in sufficient detail to catch mutual recursions
1725 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1726 subgroup is called recursively as a subpattern for a second time at the
1727 same position in the subject string. In previous releases this might have
1728 been caught by the recursion limit, or it might have run out of stack.
1730 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1731 happen only once. PCRE was, however incorrectly giving a compile time error
1732 "recursive call could loop indefinitely" because it cannot analyze the
1733 pattern in sufficient detail. The compile time test no longer happens when
1734 PCRE is compiling a conditional subpattern, but actual runaway loops are
1735 now caught at runtime (see 40 above).
1737 42. It seems that Perl allows any characters other than a closing parenthesis
1738 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1739 has been changed to be the same.
1741 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1742 as not to get warnings when autogen.sh is called. Also changed
1743 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1745 44. To help people who use pcregrep to scan files containing exceedingly long
1746 lines, the following changes have been made:
1748 (a) The default value of the buffer size parameter has been increased from
1749 8K to 20K. (The actual buffer used is three times this size.)
1751 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1752 PCRE is built.
1754 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1755 to be set at run time.
1757 (d) Numerical values in pcregrep options can be followed by K or M, for
1758 example --buffer-size=50K.
1760 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1761 given and the return code is set to 2.
1763 45. Add a pointer to the latest mark to the callout data block.
1765 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1766 partial match of an empty string instead of no match. This was specific to
1767 the use of ".".
1769 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1770 complete match instead of a partial match. This bug was dependent on both
1771 the PCRE_UTF8 and PCRE_DOTALL options being set.
1773 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1774 starting byte set, because \b was not being ignored.
1777 Version 8.12 15-Jan-2011
1778 ------------------------
1780 1. Fixed some typos in the markup of the man pages, and wrote a script that
1781 checks for such things as part of the documentation building process.
1783 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1784 --match-limit and --recursion-limit options (added for 8.11). In
1785 particular, this made one of the standard tests fail. (The integer value
1786 went into the wrong half of a long int.)
1788 3. If the --colour option was given to pcregrep with -v (invert match), it
1789 did strange things, either producing crazy output, or crashing. It should,
1790 of course, ignore a request for colour when reporting lines that do not
1791 match.
1793 4. Another pcregrep bug caused similar problems if --colour was specified with
1794 -M (multiline) and the pattern match finished with a line ending.
1796 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1797 matched in multiline mode, the following line was shown as part of the
1798 match. This seems wrong, so I have changed it.
1800 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1801 the check for further matches in the same line (so they could be coloured)
1802 to overrun the end of the current line. If another match was found, it was
1803 incorrectly shown (and then shown again when found in the next line).
1805 7. If pcregrep was compiled under Windows, there was a reference to the
1806 function pcregrep_exit() before it was defined. I am assuming this was
1807 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1808 reported by a user. I've moved the definition above the reference.
1811 Version 8.11 10-Dec-2010
1812 ------------------------
1814 1. (*THEN) was not working properly if there were untried alternatives prior
1815 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1816 backtracked to try for "b" instead of moving to the next alternative branch
1817 at the same level (in this case, to look for "c"). The Perl documentation
1818 is clear that when (*THEN) is backtracked onto, it goes to the "next
1819 alternative in the innermost enclosing group".
1821 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1822 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1823 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1824 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1825 (*THEN).
1827 3. If \s appeared in a character class, it removed the VT character from
1828 the class, even if it had been included by some previous item, for example
1829 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1830 of \s, but is part of the POSIX "space" class.)
1832 4. A partial match never returns an empty string (because you can always
1833 match an empty string at the end of the subject); however the checking for
1834 an empty string was starting at the "start of match" point. This has been
1835 changed to the "earliest inspected character" point, because the returned
1836 data for a partial match starts at this character. This means that, for
1837 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1838 (previously it gave "no match").
1840 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1841 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1842 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1843 has an implication that the given string is incomplete (because a partial
1844 match is preferred over a full match). For this reason, these items now
1845 give a partial match in this situation. [Aside: previously, the one case
1846 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1847 match rather than a full match, which was wrong by the old rules, but is
1848 now correct.]
1850 6. There was a bug in the handling of #-introduced comments, recognized when
1851 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1852 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1853 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1854 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1855 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1856 places in pcre_compile().
1858 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1859 comments when looking ahead for named forward references to subpatterns,
1860 the only newline sequence it recognized was NL. It now handles newlines
1861 according to the set newline convention.
1863 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1864 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1865 cater for a lack of strerror(). These oversights have been fixed.
1867 9. Added --match-limit and --recursion-limit to pcregrep.
1869 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1871 11. When the -o option was used, pcregrep was setting a return code of 1, even
1872 when matches were found, and --line-buffered was not being honoured.
1874 12. Added an optional parentheses number to the -o and --only-matching options
1875 of pcregrep.
1877 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1878 can match an empty string. The code to do it in pcretest and pcredemo
1879 needed fixing:
1881 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1882 only one byte after an empty string match just before CRLF (this case
1883 just got forgotten; "any" and "anycrlf" were OK).
1885 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1886 mode when an empty string match preceded an ASCII character followed by
1887 a non-ASCII character. (The code for advancing by one character rather
1888 than one byte was nonsense.)
1890 (c) The pcredemo.c sample program did not have any code at all to handle
1891 the cases when CRLF is a valid newline sequence.
1893 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1894 as a starting offset was within the subject string. There is now a new
1895 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1896 negative or greater than the length of the string. In order to test this,
1897 pcretest is extended to allow the setting of negative starting offsets.
1899 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1900 starting offset points to the beginning of a UTF-8 character was
1901 unnecessarily clumsy. I tidied it up.
1903 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1904 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1906 17. Nobody had reported that the --include_dir option, which was added in
1907 release 7.7 should have been called --include-dir (hyphen, not underscore)
1908 for compatibility with GNU grep. I have changed it to --include-dir, but
1909 left --include_dir as an undocumented synonym, and the same for
1910 --exclude-dir, though that is not available in GNU grep, at least as of
1911 release 2.5.4.
1913 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1914 characters from a string of bytes) have been redefined so as not to use
1915 loops, in order to improve performance in some environments. At the same
1916 time, I abstracted some of the common code into auxiliary macros to save
1917 repetition (this should not affect the compiled code).
1919 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1920 compile-time error is now given if \c is not followed by an ASCII
1921 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1922 different, and any byte value is allowed.)
1924 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1925 START_OPTIMIZE option, which is now allowed at compile time - but just
1926 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1927 to pcregrep and other applications that have no direct access to PCRE
1928 options. The new /Y option in pcretest sets this option when calling
1929 pcre_compile().
1931 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1932 back references. Groups containing recursive back references were forced to
1933 be atomic by that change, but in the case of named groups, the amount of
1934 memory required was incorrectly computed, leading to "Failed: internal
1935 error: code overflow". This has been fixed.
1937 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1938 pcretest.c, to avoid build problems in some Borland environments.
1941 Version 8.10 25-Jun-2010
1942 ------------------------
1944 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1945 THEN.
1947 2. (*ACCEPT) was not working when inside an atomic group.
1949 3. Inside a character class, \B is treated as a literal by default, but
1950 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1951 causes the error). The code is unchanged, but I tidied the documentation.
1953 4. Inside a character class, PCRE always treated \R and \X as literals,
1954 whereas Perl faults them if its -w option is set. I have changed PCRE so
1955 that it faults them when PCRE_EXTRA is set.
1957 5. Added support for \N, which always matches any character other than
1958 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1960 6. When compiling pcregrep with newer versions of gcc which may have
1961 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1962 declared with attribute warn_unused_result" were given. Just casting the
1963 result to (void) does not stop the warnings; a more elaborate fudge is
1964 needed. I've used a macro to implement this.
1966 7. Minor change to pcretest.c to avoid a compiler warning.
1968 8. Added four artifical Unicode properties to help with an option to make
1969 \s etc use properties (see next item). The new properties are: Xan
1970 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1972 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1973 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1974 this option. Modified pcretest to add /W to test this facility. Added
1975 REG_UCP to make it available via the POSIX interface.
1977 10. Added --line-buffered to pcregrep.
1979 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1980 studied, and the match started with a letter with a code point greater than
1981 127 whose first byte was different to the first byte of the other case of
1982 the letter, the other case of this starting letter was not recognized
1983 (#976).
1985 12. If a pattern that was studied started with a repeated Unicode property
1986 test, for example, \p{Nd}+, there was the theoretical possibility of
1987 setting up an incorrect bitmap of starting bytes, but fortunately it could
1988 not have actually happened in practice until change 8 above was made (it
1989 added property types that matched character-matching opcodes).
1991 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1992 possible starting bytes for non-anchored patterns.
1994 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1995 \R, and also a number of cases that involve Unicode properties, both
1996 explicit and implicit when PCRE_UCP is set.
1998 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1999 input, it could crash or give wrong results if characters with values
2000 greater than 0xc0 were present in the subject string. (Detail: it assumed
2001 UTF-8 input when processing these items.)
2003 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2004 size_t is 64-bit (#991).
2006 17. Added a check for running out of memory when PCRE is compiled with
2007 --disable-stack-for-recursion (#990).
2009 18. If the last data line in a file for pcretest does not have a newline on
2010 the end, a newline was missing in the output.
2012 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2013 less than 128) in its various bitmaps. However, there is a facility for
2014 generating tables according to the current locale when PCRE is compiled. It
2015 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2016 characters, are recognized by isspace() and therefore were getting set in
2017 these tables, and indeed these tables seem to approximate to ISO 8859. This
2018 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2019 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2020 which of course cannot start UTF-8 characters. I have changed the code so
2021 that only real ASCII characters (less than 128) and the correct starting
2022 bytes for UTF-8 encodings are set for characters greater than 127 when in
2023 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2024 altogether.)
2026 20. Added the /T option to pcretest so as to be able to run tests with non-
2027 standard character tables, thus making it possible to include the tests
2028 used for 19 above in the standard set of tests.
2030 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2031 reference to a subpattern the other side of a comment that contains an
2032 opening parenthesis caused either an internal compiling error, or a
2033 reference to the wrong subpattern.
2036 Version 8.02 19-Mar-2010
2037 ------------------------
2039 1. The Unicode data tables have been updated to Unicode 5.2.0.
2041 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2042 configured.
2044 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2045 original author of that file, following a query about its status.
2047 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2048 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2050 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2051 quantifier applied to a forward-referencing subroutine call, could compile
2052 incorrect code or give the error "internal error: previously-checked
2053 referenced subpattern not found".
2055 6. Both MS Visual Studio and Symbian OS have problems with initializing
2056 variables to point to external functions. For these systems, therefore,
2057 pcre_malloc etc. are now initialized to local functions that call the
2058 relevant global functions.
2060 7. There were two entries missing in the vectors called coptable and poptable
2061 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2062 I've fixed the data, and added a kludgy way of testing at compile time that
2063 the lengths are correct (equal to the number of opcodes).
2065 8. Following on from 7, I added a similar kludge to check the length of the
2066 eint vector in pcreposix.c.
2068 9. Error texts for pcre_compile() are held as one long string to avoid too
2069 much relocation at load time. To find a text, the string is searched,
2070 counting zeros. There was no check for running off the end of the string,
2071 which could happen if a new error number was added without updating the
2072 string.
2074 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2076 11. \K was not working if it appeared in an atomic group or in a group that
2077 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2078 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2079 the assertion is positive, but not if it is negative.
2081 12. Change 11 fortuitously reduced the size of the stack frame used in the
2082 "match()" function of pcre_exec.c by one pointer. Forthcoming
2083 implementation of support for (*MARK) will need an extra pointer on the
2084 stack; I have reserved it now, so that the stack frame size does not
2085 decrease.
2087 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2088 item in branch that calls a recursion is a subroutine call - as in the
2089 second branch in the above example - was incorrectly given the compile-
2090 time error "recursive call could loop indefinitely" because pcre_compile()
2091 was not correctly checking the subroutine for matching a non-empty string.
2093 14. The checks for overrunning compiling workspace could trigger after an
2094 overrun had occurred. This is a "should never occur" error, but it can be
2095 triggered by pathological patterns such as hundreds of nested parentheses.
2096 The checks now trigger 100 bytes before the end of the workspace.
2098 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2101 Version 8.01 19-Jan-2010
2102 ------------------------
2104 1. If a pattern contained a conditional subpattern with only one branch (in
2105 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2106 computed the wrong minimum data length (which is of course zero for such
2107 subpatterns). This could cause incorrect "no match" results.
2109 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2110 the pattern is reset in the first branch, pcre_compile() failed with
2111 "internal error: code overflow at offset...". This happened only when
2112 the reset was to the original external option setting. (An optimization
2113 abstracts leading options settings into an external setting, which was the
2114 cause of this.)
2116 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2117 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2118 assertion pattern did not match (meaning that the assertion was true), it
2119 was incorrectly treated as false if the SKIP had been reached during the
2120 matching. This also applied to assertions used as conditions.
2122 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2123 assertion subpattern, including such a pattern used as a condition,
2124 unpredictable results occurred, instead of the error return
2127 5. The C++ GlobalReplace function was not working like Perl for the special
2128 situation when an empty string is matched. It now does the fancy magic
2129 stuff that is necessary.
2131 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2132 removed. (These were left over from very, very early versions of PCRE.)
2134 7. Some cosmetic changes to the code to make life easier when compiling it
2135 as part of something else:
2137 (a) Change DEBUG to PCRE_DEBUG.
2139 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2140 called "current" as "current_branch", to prevent a collision with the
2141 Linux macro when compiled as a kernel module.
2143 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2144 prevent a collision with the Linux macro when compiled as a kernel
2145 module.
2147 8. In pcre_compile() there are some checks for integer overflows that used to
2148 cast potentially large values to (double). This has been changed to that
2149 when building, a check for int64_t is made, and if it is found, it is used
2150 instead, thus avoiding the use of floating point arithmetic. (There is no
2151 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2152 double.
2154 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2155 2005 (difference between two addresses compared to an unsigned value).
2157 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2158 custom one, because of the following reported problem in Windows:
2160 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2161 under Win32.
2162 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2163 therefore missing the function definition.
2164 - The compiler thus generates a "C" signature for the test function.
2165 - The linker fails to find the "C" function.
2166 - PCRE fails to configure if asked to do so against libbz2.
2168 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2169 messages were output:
2171 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2172 rerunning libtoolize, to keep the correct libtool macros in-tree.
2173 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2175 I have done both of these things.
2177 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2178 most of the time, it *can* run out if it is given a pattern that contains a
2179 runaway infinite recursion. I updated the discussion in the pcrestack man
2180 page.
2182 13. Now that we have gone to the x.xx style of version numbers, the minor
2183 version may start with zero. Using 08 or 09 is a bad idea because users
2184 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2185 interpreted as invalid octal numbers. I've updated the previous comment in
2186 configure.ac, and also added a check that gives an error if 08 or 09 are
2187 used.
2189 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2190 causing partial matching to fail when the end of the subject matched \W
2191 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2193 15. There were some discrepancies between the declarations in pcre_internal.h
2194 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2195 their definitions. The declarations used "const uschar *" and the
2196 definitions used USPTR. Even though USPTR is normally defined as "const
2197 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2198 reported that: "This difference in casting confuses some C++ compilers, for
2199 example, SunCC recognizes above declarations as different functions and
2200 generates broken code for hbpcre." I have changed the declarations to use
2201 USPTR.
2203 16. GNU libtool is named differently on some systems. The autogen.sh script now
2204 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2205 (FreeBSD).
2207 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2208 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2209 comment: "Figure out how to create a longlong from a string: strtoll and
2210 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2211 instance, but it only takes 2 args instead of 3!"
2213 18. A subtle bug concerned with back references has been fixed by a change of
2214 specification, with a corresponding code fix. A pattern such as
2215 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2216 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2217 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2218 same bug. Such groups have to be quantified to be useful, or contained
2219 inside another quantified group. (If there's no repetition, the reference
2220 can never match.) The problem arises because, having left the group and
2221 moved on to the rest of the pattern, a later failure that backtracks into
2222 the group uses the captured value from the final iteration of the group
2223 rather than the correct earlier one. I have fixed this in PCRE by forcing
2224 any group that contains a reference to itself to be an atomic group; that
2225 is, there cannot be any backtracking into it once it has completed. This is
2226 similar to recursive and subroutine calls.
2229 Version 8.00 19-Oct-09
2230 ----------------------
2232 1. The table for translating pcre_compile() error codes into POSIX error codes
2233 was out-of-date, and there was no check on the pcre_compile() error code
2234 being within the table. This could lead to an OK return being given in
2235 error.
2237 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2238 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2239 in a Windows environment.
2241 3. The pcregrep --count option prints the count for each file even when it is
2242 zero, as does GNU grep. However, pcregrep was also printing all files when
2243 --files-with-matches was added. Now, when both options are given, it prints
2244 counts only for those files that have at least one match. (GNU grep just
2245 prints the file name in this circumstance, but including the count seems
2246 more useful - otherwise, why use --count?) Also ensured that the
2247 combination -clh just lists non-zero counts, with no names.
2249 4. The long form of the pcregrep -F option was incorrectly implemented as
2250 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2251 but it seems right to fix it, and I didn't think it was worth preserving
2252 the old behaviour.
2254 5. The command line items --regex=pattern and --regexp=pattern were not
2255 recognized by pcregrep, which required --regex pattern or --regexp pattern
2256 (with a space rather than an '='). The man page documented the '=' forms,
2257 which are compatible with GNU grep; these now work.
2259 6. No libpcreposix.pc file was created for pkg-config; there was just
2260 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2262 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2263 when UCP support is not needed, by modifying the Python script that
2264 generates it from Unicode data files. This should not matter if the module
2265 is correctly used as a library, but I received one complaint about 50K of
2266 unwanted data. My guess is that the person linked everything into his
2267 program rather than using a library. Anyway, it does no harm.
2269 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2270 was a minimum greater than 1 for a wide character in a possessive
2271 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2272 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2273 character. Chaos in the form of incorrect output or a compiling loop could
2274 result.
2276 9. The restrictions on what a pattern can contain when partial matching is
2277 requested for pcre_exec() have been removed. All patterns can now be
2278 partially matched by this function. In addition, if there are at least two
2279 slots in the offset vector, the offset of the earliest inspected character
2280 for the match and the offset of the end of the subject are set in them when
2281 PCRE_ERROR_PARTIAL is returned.
2283 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2284 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2285 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2286 and may be more useful for multi-segment matching.
2288 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2289 used to be given if ever the end of the subject was reached; now it is
2290 given only if matching could not proceed because another character was
2291 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2292 string "Z", which now yields "no match" instead of "partial match". In the
2293 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2294 final character ended with (*FAIL).
2296 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2297 if the pattern had a "must contain" character that was already found in the
2298 earlier partial match, unless partial matching was again requested. For
2299 example, with the pattern /dog.(body)?/, the "must contain" character is
2300 "g". If the first part-match was for the string "dog", restarting with
2301 "sbody" failed. This bug has been fixed.
2303 13. The string returned by pcre_dfa_exec() after a partial match has been
2304 changed so that it starts at the first inspected character rather than the
2305 first character of the match. This makes a difference only if the pattern
2306 starts with a lookbehind assertion or \b or \B (\K is not supported by
2307 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2308 matching functions compatible, and I think it's the right thing to do.
2310 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2311 so that the demonstration program is easily available in environments where
2312 PCRE has not been installed from source.
2314 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2315 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2316 library.
2318 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2319 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2320 is not the first non-POSIX option to be added. Clearly some people find
2321 these options useful.
2323 17. If a caller to the POSIX matching function regexec() passes a non-zero
2324 value for nmatch with a NULL value for pmatch, the value of
2325 nmatch is forced to zero.
2327 18. RunGrepTest did not have a test for the availability of the -u option of
2328 the diff command, as RunTest does. It now checks in the same way as
2329 RunTest, and also checks for the -b option.
2331 19. If an odd number of negated classes containing just a single character
2332 interposed, within parentheses, between a forward reference to a named
2333 subpattern and the definition of the subpattern, compilation crashed with
2334 an internal error, complaining that it could not find the referenced
2335 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2336 [The bug was that it was starting one character too far in when skipping
2337 over the character class, thus treating the ] as data rather than
2338 terminating the class. This meant it could skip too much.]
2340 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2341 /g option in pcretest when the pattern contains \K, which makes it possible
2342 to have an empty string match not at the start, even when the pattern is
2343 anchored. Updated pcretest and pcredemo to use this option.
2345 21. If the maximum number of capturing subpatterns in a recursion was greater
2346 than the maximum at the outer level, the higher number was returned, but
2347 with unset values at the outer level. The correct (outer level) value is
2348 now given.
2350 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2351 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2352 make it do so. The string so far is captured, making this feature
2353 compatible with Perl.
2355 23. The tests have been re-organized, adding tests 11 and 12, to make it
2356 possible to check the Perl 5.10 features against Perl 5.10.
2358 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2359 pattern matches a fixed length string. PCRE did not allow this; now it
2360 does. Neither allows recursion.
2362 25. I finally figured out how to implement a request to provide the minimum
2363 length of subject string that was needed in order to match a given pattern.
2364 (It was back references and recursion that I had previously got hung up
2365 on.) This code has now been added to pcre_study(); it finds a lower bound
2366 to the length of subject needed. It is not necessarily the greatest lower
2367 bound, but using it to avoid searching strings that are too short does give
2368 some useful speed-ups. The value is available to calling programs via
2369 pcre_fullinfo().
2371 26. While implementing 25, I discovered to my embarrassment that pcretest had
2372 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2373 study optimizations had never been tested with that matching function.
2374 Oops. What is worse, even when it was passed study data, there was a bug in
2375 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2376 were also very few tests of studied patterns with pcre_dfa_exec().
2378 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2379 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2380 on the other side of the coin, they are no longer allowed to have different
2381 names, because these cannot be distinguished in PCRE, and this has caused
2382 confusion. (This is a difference from Perl.)
2384 28. When duplicate subpattern names are present (necessarily with different
2385 numbers, as required by 27 above), and a test is made by name in a
2386 conditional pattern, either for a subpattern having been matched, or for
2387 recursion in such a pattern, all the associated numbered subpatterns are
2388 tested, and the overall condition is true if the condition is true for any
2389 one of them. This is the way Perl works, and is also more like the way
2390 testing by number works.
2393 Version 7.9 11-Apr-09
2394 ---------------------
2396 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2397 (pcretest), all targets were linked against these libraries. This included
2398 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2399 libraries. This caused unwanted dependencies to be created. This problem
2400 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2401 pcretest is linked with readline.
2403 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2404 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2405 moved outside it again, because FALSE and TRUE are already defined in AIX,
2406 but BOOL is not.
2408 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2409 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2411 4. The pcregrep documentation said spaces were inserted as well as colons (or
2412 hyphens) following file names and line numbers when outputting matching
2413 lines. This is not true; no spaces are inserted. I have also clarified the
2414 wording for the --colour (or --color) option.
2416 5. In pcregrep, when --colour was used with -o, the list of matching strings
2417 was not coloured; this is different to GNU grep, so I have changed it to be
2418 the same.
2420 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2421 each matching line was coloured. Now it goes on to look for further matches
2422 of any of the test patterns, which is the same behaviour as GNU grep.
2424 7. A pattern that could match an empty string could cause pcregrep to loop; it
2425 doesn't make sense to accept an empty string match in pcregrep, so I have
2426 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2427 seems to be how GNU grep behaves. [But see later change 40 for release
2428 8.33.]
2430 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2431 start or after a newline", because the conditional assertion was not being
2432 correctly handled. The rule now is that both the assertion and what follows
2433 in the first alternative must satisfy the test.
2435 9. If auto-callout was enabled in a pattern with a conditional group whose
2436 condition was an assertion, PCRE could crash during matching, both with
2437 pcre_exec() and pcre_dfa_exec().
2439 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2440 used for matching.
2442 11. Unicode property support in character classes was not working for
2443 characters (bytes) greater than 127 when not in UTF-8 mode.
2445 12. Added the -M command line option to pcretest.
2447 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2449 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2451 16. Added comments and documentation about mis-use of no_arg in the C++
2452 wrapper.
2454 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2455 from Martin Jerabek that uses macro names for all relevant character and
2456 string constants.
2458 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2459 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2460 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2461 these, but not everybody uses configure.
2463 19. A conditional group that had only one branch was not being correctly
2464 recognized as an item that could match an empty string. This meant that an
2465 enclosing group might also not be so recognized, causing infinite looping
2466 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2467 with the subject "ab", where knowledge that the repeated group can match
2468 nothing is needed in order to break the loop.
2470 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2471 exec(), but without supplying a callout function, matching went wrong.
2473 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2474 leak if the size of the offset vector was greater than 30. When the vector
2475 is smaller, the saved offsets during recursion go onto a local stack
2476 vector, but for larger vectors malloc() is used. It was failing to free
2477 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2478 error, in fact).
2480 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2481 heapframe that is used only when UTF-8 support is enabled. This caused no
2482 problem, but was untidy.
2484 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2485 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2486 included within another project.
2488 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2489 slightly modified by me:
2491 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2492 not building pcregrep.
2494 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2495 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2497 25. Forward references, both numeric and by name, in patterns that made use of
2498 duplicate group numbers, could behave incorrectly or give incorrect errors,
2499 because when scanning forward to find the reference group, PCRE was not
2500 taking into account the duplicate group numbers. A pattern such as
2501 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2503 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2504 the feature of a custom pointer more persuasive (as requested by a user).
2506 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2507 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2508 pre-defined.
2510 28. Added support for (*UTF8) at the start of a pattern.
2512 29. Arrange for flags added by the "release type" setting in CMake to be shown
2513 in the configuration summary.
2516 Version 7.8 05-Sep-08
2517 ---------------------
2519 1. Replaced UCP searching code with optimized version as implemented for Ad
2520 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2521 stage table and inline lookup instead of a function, giving speed ups of 2
2522 to 5 times on some simple patterns that I tested. Permission was given to
2523 distribute the MultiStage2.py script that generates the tables (it's not in
2524 the tarball, but is in the Subversion repository).
2526 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2527 scripts.
2529 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2530 a group with a zero qualifier. The result of the study could be incorrect,
2531 or the function might crash, depending on the pattern.
2533 4. Caseless matching was not working for non-ASCII characters in back
2534 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2535 It now works when Unicode Property Support is available.
2537 5. In pcretest, an escape such as \x{de} in the data was always generating
2538 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2539 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2540 truncation.
2542 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2544 7. Added two (int) casts to pcregrep when printing the difference of two
2545 pointers, in case they are 64-bit values.
2547 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2548 test 2 if it fails.
2550 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2551 and a #define of that name to empty if it is not externally set. This is to
2552 allow users of MSVC to set it if necessary.
2554 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2555 the convenience functions in the pcre_get.c source file.
2557 11. An option change at the start of a pattern that had top-level alternatives
2558 could cause overwriting and/or a crash. This command provoked a crash in
2559 some environments:
2561 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2563 This potential security problem was recorded as CVE-2008-2371.
2565 12. For a pattern where the match had to start at the beginning or immediately
2566 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2567 pcre_dfa_exec() could read past the end of the passed subject if there was
2568 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2569 pcretest so that it places the subject at the end of its malloc-ed buffer.
2571 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2572 exec() might read past the end of the data buffer in UTF-8 mode.
2574 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2575 the data contained the byte 0x85 as part of a UTF-8 character within its
2576 first line. This applied both to normal and DFA matching.
2578 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2579 /^[^d]*?$/8 failed to match "abc".
2581 16. Added a missing copyright notice to pcrecpp_internal.h.
2583 17. Make it more clear in the documentation that values returned from
2584 pcre_exec() in ovector are byte offsets, not character counts.
2586 18. Tidied a few places to stop certain compilers from issuing warnings.
2588 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2589 supplied by Stefan Weber. I made a further small update for 7.8 because
2590 there is a change of source arrangements: the pcre_searchfuncs.c module is
2591 replaced by pcre_ucd.c.
2594 Version 7.7 07-May-08
2595 ---------------------
2597 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2598 a string to a long long, pretend we don't even have a long long." This is
2599 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2601 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2602 pre-7.6 versions, which defined a global no_arg variable instead of putting
2603 it in the RE class. (See also #8 below.)
2605 3. Remove a line of dead code, identified by coverity and reported by Nuno
2606 Lopes.
2608 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2610 (1) The include/exclude patterns were being applied to the whole pathnames
2611 of files, instead of just to the final components.
2613 (2) If there was more than one level of directory, the subdirectories were
2614 skipped unless they satisfied the include/exclude conditions. This is
2615 inconsistent with GNU grep (and could even be seen as contrary to the
2616 pcregrep specification - which I improved to make it absolutely clear).
2617 The action now is always to scan all levels of directory, and just
2618 apply the include/exclude patterns to regular files.
2620 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2621 --exclude_dir in the tests to avoid scanning .svn directories.
2623 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2624 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2625 doesn't support NULs in patterns.
2627 7. Added some missing "const"s to declarations of static tables in
2628 pcre_compile.c and pcre_dfa_exec.c.
2630 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2631 caused by fix #2 above. (Subsequently also a second patch to fix the
2632 first patch. And a third patch - this was a messy problem.)
2634 9. Applied Craig's patch to remove the use of push_back().
2636 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2637 matching function regexec().
2639 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2640 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2641 references. PCRE supports relative numbers with this syntax (I don't think
2642 Oniguruma does).
2644 12. Previously, a group with a zero repeat such as (...){0} was completely
2645 omitted from the compiled regex. However, this means that if the group
2646 was called as a subroutine from elsewhere in the pattern, things went wrong
2647 (an internal error was given). Such groups are now left in the compiled
2648 pattern, with a new opcode that causes them to be skipped at execution
2649 time.
2651 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2652 to the way PCRE behaves:
2654 (a) A lone ] character is dis-allowed (Perl treats it as data).
2656 (b) A back reference to an unmatched subpattern matches an empty string
2657 (Perl fails the current match path).
2659 (c) A data ] in a character class must be notated as \] because if the
2660 first data character in a class is ], it defines an empty class. (In
2661 Perl it is not possible to have an empty class.) The empty class []
2662 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2663 The negative empty class [^] matches any one character, independently
2664 of the DOTALL setting.
2666 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2667 non-existent subpattern following a character class starting with ']' and
2668 containing () gave an internal compiling error instead of "reference to
2669 non-existent subpattern". Fortunately, when the pattern did exist, the
2670 compiled code was correct. (When scanning forwards to check for the
2671 existence of the subpattern, it was treating the data ']' as terminating
2672 the class, so got the count wrong. When actually compiling, the reference
2673 was subsequently set up correctly.)
2675 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2676 it was being rejected as not supported by pcre_dfa_exec(), even though
2677 other assertions are supported. I have made pcre_dfa_exec() support
2678 (*FAIL).
2680 16. The implementation of 13c above involved the invention of a new opcode,
2681 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2682 cannot be changed at match time, I realized I could make a small
2683 improvement to matching performance by compiling OP_ALLANY instead of
2684 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2685 on the OP_ANY path.
2687 17. Compiling pcretest on Windows with readline support failed without the
2688 following two fixes: (1) Make the unistd.h include conditional on
2689 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2691 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2692 ncurses library to be included for pcretest when ReadLine support is
2693 requested, but also to allow for it to be overridden. This patch came from
2694 Daniel Bergström.
2696 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2697 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2698 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2699 spotting this.
2702 Version 7.6 28-Jan-08
2703 ---------------------
2705 1. A character class containing a very large number of characters with
2706 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2707 overflow.
2709 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2710 HAVE_LONG_LONG is not defined.
2712 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2713 bring them up to date and include new features. This patch includes:
2715 - Fixed PH's badly added libz and libbz2 support.
2716 - Fixed a problem with static linking.
2717 - Added pcredemo. [But later removed - see 7 below.]
2718 - Fixed dftables problem and added an option.
2719 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2721 - Added readline support for pcretest.
2722 - Added an listing of the option settings after cmake has run.
2724 4. A user submitted a patch to Makefile that makes it easy to create
2725 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2726 Makefile.am that cause it to include this special target, without
2727 affecting anything else. Note that the same mingw target plus all
2728 the other distribution libraries and programs are now supported
2729 when configuring with CMake (see 6 below) instead of with
2730 Configure/Make.
2732 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2733 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2734 exported in the Windows port". It has not yet been confirmed that the patch
2735 solves the problem, but it does no harm.
2737 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2738 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2739 with CMake, and also correct the comment about stack recursion.
2741 7. Remove the automatic building of pcredemo from the ./configure system and
2742 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2743 of a program that users should build themselves after PCRE is installed, so
2744 building it automatically is not really right. What is more, it gave
2745 trouble in some build environments.
2747 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2750 Version 7.5 10-Jan-08
2751 ---------------------
2753 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2754 values in parens when parsing an RE using the C++ wrapper."
2756 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2757 Characters greater than 255 were excluded from the class instead of being
2758 included.
2760 3. The same bug as (2) above applied to negated POSIX classes such as
2761 [:^space:].
2763 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2764 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2765 I have changed it.
2767 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2768 first named subpattern) and a construct such as (?&a) would reference the
2769 first named subpattern whose name started with "a" (in other words, the
2770 length check was missing). Both these problems are fixed. "Subpattern name
2771 expected" is now given for (?&) (a zero-length name), and this patch also
2772 makes it give the same error for \k'' (previously it complained that that
2773 was a reference to a non-existent subpattern).
2775 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2776 this is right because (?- can be followed by option settings as well as by
2777 digits. I have, however, made the messages clearer.
2779 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2780 than the number used in the conditional) now cause a compile-time error.
2781 This is actually not compatible with Perl, which accepts such patterns, but
2782 treats the conditional as always being FALSE (as PCRE used to), but it
2783 seems to me that giving a diagnostic is better.
2785 8. Change "alphameric" to the more common word "alphanumeric" in comments
2786 and messages.
2788 9. Fix two occurrences of "backslash" in comments that should have been
2789 "backspace".
2791 10. Remove two redundant lines of code that can never be obeyed (their function
2792 was moved elsewhere).
2794 11. The program that makes PCRE's Unicode character property table had a bug
2795 which caused it to generate incorrect table entries for sequences of
2796 characters that have the same character type, but are in different scripts.
2797 It amalgamated them into a single range, with the script of the first of
2798 them. In other words, some characters were in the wrong script. There were
2799 thirteen such cases, affecting characters in the following ranges:
2801 U+002b0 - U+002c1
2802 U+0060c - U+0060d
2803 U+0061e - U+00612
2804 U+0064b - U+0065e
2805 U+0074d - U+0076d
2806 U+01800 - U+01805
2807 U+01d00 - U+01d77
2808 U+01d9b - U+01dbf
2809 U+0200b - U+0200f
2810 U+030fc - U+030fe
2811 U+03260 - U+0327f
2812 U+0fb46 - U+0fbb1
2813 U+10450 - U+1049d
2815 12. The -o option (show only the matching part of a line) for pcregrep was not
2816 compatible with GNU grep in that, if there was more than one match in a
2817 line, it showed only the first of them. It now behaves in the same way as
2818 GNU grep.
2820 13. If the -o and -v options were combined for pcregrep, it printed a blank
2821 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2822 does the same. The return code can be used to tell if there were any
2823 non-matching lines.
2825 14. Added --file-offsets and --line-offsets to pcregrep.
2827 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2828 infinitely looping recursion. The bug was that positive lookaheads were not
2829 being skipped when checking for a possible empty match (negative lookaheads
2830 and both kinds of lookbehind were skipped).
2832 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2833 inclusion of <windows.h> to before rather than after the definition of
2834 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2836 17. Specifying a possessive quantifier with a specific limit for a Unicode
2837 character property caused pcre_compile() to compile bad code, which led at
2838 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2839 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2840 caused the error; without that there was no problem.
2842 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2844 19. Added --enable-pcretest-libreadline.
2846 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2847 RE::GlobalReplace(). As a result, the number of replacements returned was
2848 double what it should be. I removed one of the increments, but Craig sent a
2849 later patch that removed the other one (the right fix) and added unit tests
2850 that check the return values (which was not done before).
2852 21. Several CMake things:
2854 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2855 the names libpcre and libpcreposix, not just pcre and pcreposix.
2857 (2) The above change means that pcretest and pcregrep are now correctly
2858 linked with the newly-built libraries, not previously installed ones.
2862 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2863 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2864 UTF-8 newline character). The key issue is that the pattern starts .*;
2865 this means that the match must be either at the beginning, or after a
2866 newline. The bug was in the code for advancing after a failed match and
2867 checking that the new position followed a newline. It was not taking
2868 account of UTF-8 characters correctly.
2870 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2871 character classes. PCRE was not treating the sequence [:...:] as a
2872 character class unless the ... were all letters. Perl, however, seems to
2873 allow any characters between [: and :], though of course it rejects as
2874 unknown any "names" that contain non-letters, because all the known class
2875 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2876 for example, whereas PCRE did not - it did not recognize a POSIX character
2877 class. This seemed a bit dangerous, so the code has been changed to be
2878 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2879 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2880 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2881 Perl does, and where it didn't before.
2883 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2884 Windows environments %n is disabled by default.
2887 Version 7.4 21-Sep-07
2888 ---------------------
2890 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2891 means that a class such as [\s] counted as "explicit reference to CR or
2892 LF". That isn't really right - the whole point of the change was to try to
2893 help when there was an actual mention of one of the two characters. So now
2894 the change happens only if \r or \n (or a literal CR or LF) character is
2895 encountered.
2897 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2898 of both had grown to the point where there were only 3 bits left.
2899 Fortunately, there was spare space in the data structure, and so I have
2900 moved the internal flags into a new 16-bit field to free up more option
2901 bits.
2903 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2904 but did not set the internal JCHANGED flag - either of these is enough to
2905 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2906 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2907 start sets both bits.
2909 4. Added options (at build time, compile time, exec time) to change \R from
2910 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2912 5. doc/pcresyntax.html was missing from the distribution.
2914 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2915 compatibility, even though it is no longer used.
2917 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2918 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2919 windows.h file is present (where different names are used). [This was
2920 reversed later after testing - see 16 below.]
2922 8. Changed all #include <config.h> to #include "config.h". There were also
2923 some further <pcre.h> cases that I changed to "pcre.h".
2925 9. When pcregrep was used with the --colour option, it missed the line ending
2926 sequence off the lines that it output.
2928 10. It was pointed out to me that arrays of string pointers cause lots of
2929 relocations when a shared library is dynamically loaded. A technique of
2930 using a single long string with a table of offsets can drastically reduce
2931 these. I have refactored PCRE in four places to do this. The result is
2932 dramatic:
2934 Originally: 290
2935 After changing UCP table: 187
2936 After changing error message table: 43
2937 After changing table of "verbs" 36
2938 After changing table of Posix names 22
2940 Thanks to the folks working on Gregex for glib for this insight.
2942 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2943 unicode-properties was also set.
2945 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2947 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2948 checked only for CRLF.
2950 14. Added casts to pcretest.c to avoid compiler warnings.
2952 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2954 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2955 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2956 entirely. This removes changes made in 7 above.
2958 17. The CMake files have been updated, and there is now more information about
2959 building with CMake in the NON-UNIX-USE document.
2962 Version 7.3 28-Aug-07
2963 ---------------------
2965 1. In the rejigging of the build system that eventually resulted in 7.1, the
2966 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2967 brackets there is not right, since it causes compilers to look for an
2968 installed pcre.h, not the version that is in the source that is being
2969 compiled (which of course may be different). I have changed it back to:
2971 #include "pcre.h"
2973 I have a vague recollection that the change was concerned with compiling in
2974 different directories, but in the new build system, that is taken care of
2975 by the VPATH setting the Makefile.
2977 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2978 when the subject happened to end in the byte 0x85 (e.g. if the last
2979 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2980 characters but of course it shouldn't be taken as a newline when it is part
2981 of another character. The bug was that, for an unlimited repeat of . in
2982 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2983 characters when looking for a newline.
2985 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2987 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2988 in debug output.
2990 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2991 long printing in the pcrecpp unittest when running under MinGW.
2993 6. ESC_K was left out of the EBCDIC table.
2995 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2996 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2997 limit also applies to "virtual nesting" when a pattern is recursive, and in
2998 this case 1000 isn't so big. I have been able to remove this limit at the
2999 expense of backing off one optimization in certain circumstances. Normally,
3000 when pcre_exec() would call its internal match() function recursively and
3001 immediately return the result unconditionally, it uses a "tail recursion"
3002 feature to save stack. However, when a subpattern that can match an empty
3003 string has an unlimited repetition quantifier, it no longer makes this
3004 optimization. That gives it a stack frame in which to save the data for
3005 checking that an empty string has been matched. Previously this was taken
3006 from the 1000-entry workspace that had been reserved. So now there is no
3007 explicit limit, but more stack is used.
3009 8. Applied Daniel's patches to solve problems with the import/export magic
3010 syntax that is required for Windows, and which was going wrong for the
3011 pcreposix and pcrecpp parts of the library. These were overlooked when this
3012 problem was solved for the main library.
3014 9. There were some crude static tests to avoid integer overflow when computing
3015 the size of patterns that contain repeated groups with explicit upper
3016 limits. As the maximum quantifier is 65535, the maximum group length was
3017 set at 30,000 so that the product of these two numbers did not overflow a
3018 32-bit integer. However, it turns out that people want to use groups that
3019 are longer than 30,000 bytes (though not repeat them that many times).
3020 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3021 made it possible to implement the integer overflow checks in a much more
3022 dynamic way, which I have now done. The artificial limitation on group
3023 length has been removed - we now have only the limit on the total length of
3024 the compiled pattern, which depends on the LINK_SIZE setting.
3026 10. Fixed a bug in the documentation for get/copy named substring when
3027 duplicate names are permitted. If none of the named substrings are set, the
3028 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3029 empty string.
3031 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3032 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3033 because the ] is interpreted as the first data character and the
3034 terminating ] is not found. PCRE has been made compatible with Perl in this
3035 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3036 cause memory overwriting.
3038 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3039 string has been matched (to stop an infinite loop). It was not recognizing
3040 a conditional subpattern that could match an empty string if that
3041 subpattern was within another subpattern. For example, it looped when
3042 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3043 condition was not nested. This bug has been fixed.
3045 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3046 past the start of the subject in the presence of bytes with the top bit
3047 set, for example "\x8aBCD".
3049 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3050 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3052 14. Optimized (?!) to (*FAIL).
3054 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3055 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3056 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3057 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3058 does: it's just the validity check that is more restrictive.
3060 16. Inserted checks for integer overflows during escape sequence (backslash)
3061 processing, and also fixed erroneous offset values for syntax errors during
3062 backslash processing.
3064 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3065 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3067 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3068 caused an overrun.
3070 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3071 something other than just ASCII characters) inside a group that had an
3072 unlimited repeat caused a loop at compile time (while checking to see
3073 whether the group could match an empty string).
3075 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3076 [\P{Any}] did so. (Error in the code for printing property names.)
3078 21. An orphan \E inside a character class could cause a crash.
3080 22. A repeated capturing bracket such as (A)? could cause a wild memory
3081 reference during compilation.
3083 23. There are several functions in pcre_compile() that scan along a compiled
3084 expression for various reasons (e.g. to see if it's fixed length for look
3085 behind). There were bugs in these functions when a repeated \p or \P was
3086 present in the pattern. These operators have additional parameters compared
3087 with \d, etc, and these were not being taken into account when moving along
3088 the compiled data. Specifically:
3090 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3091 length.
3093 (b) An item such as \pL+ within a repeated group could cause crashes or
3094 loops.
3096 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3097 "reference to non-existent subpattern" error.
3099 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3101 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3102 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3104 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3106 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3107 character were causing crashes (broken optimization).
3109 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3110 \p or \P) caused a compile-time loop.
3112 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3113 break. For example, the unstudied pattern [\r\n]A does not match the string
3114 "\r\nA" because change 7.0/46 below moves the current point on by two
3115 characters after failing to match at the start. However, the pattern \nA
3116 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3117 the same is true. There doesn't seem any very clean way out of this, but
3118 what I have chosen to do makes the common cases work: PCRE now takes note
3119 of whether there can be an explicit match for \r or \n anywhere in the
3120 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3121 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3122 pattern has explicit CR or LF references.
3124 29. Added (*CR) etc for changing newline setting at start of pattern.
3127 Version 7.2 19-Jun-07
3128 ---------------------
3130 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3131 which is apparently normally available under Windows.
3133 2. Re-jig the pcregrep tests with different newline settings in an attempt
3134 to make them independent of the local environment's newline setting.
3136 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3138 4. Some of the "internals" tests were previously cut out when the link size
3139 was not 2, because the output contained actual offsets. The recent new
3140 "Z" feature of pcretest means that these can be cut out, making the tests
3141 usable with all link sizes.
3143 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3144 stack recursion. This gives a massive performance boost under BSD, but just
3145 a small improvement under Linux. However, it saves one field in the frame
3146 in all cases.
3148 6. Added more features from the forthcoming Perl 5.10:
3150 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3151 recursion call. It refers to the nth most recently opened parentheses.
3153 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3154 to be opened parentheses.
3156 (c) Conditions that refer to capturing parentheses can be specified
3157 relatively, for example, (?(-2)... or (?(+3)...
3159 (d) \K resets the start of the current match so that everything before
3160 is not part of it.
3162 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3164 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3165 reference syntax.
3167 (g) (?| introduces a group in which the numbering of parentheses in each
3168 alternative starts with the same number.
3170 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3172 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3175 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3176 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3177 for detecting groups that can match an empty string.
3179 9. A pattern with a very large number of alternatives (more than several
3180 hundred) was running out of internal workspace during the pre-compile
3181 phase, where pcre_compile() figures out how much memory will be needed. A
3182 bit of new cunning has reduced the workspace needed for groups with
3183 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3184 workspace instead of running out of the 4096 that are available.
3186 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3188 11. Applied patch from Google to remove an optimization that didn't quite work.
3189 The report of the bug said:
3191 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3192 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3193 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3195 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3196 it matched the wrong number of bytes.
3199 Version 7.1 24-Apr-07
3200 ---------------------
3202 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3203 that is more "standard", making use of automake and other Autotools. There
3204 is some re-arrangement of the files and adjustment of comments consequent
3205 on this.
3207 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3208 for recursive directory scanning broke on some systems because the files
3209 are not scanned in any specific order and on different systems the order
3210 was different. A call to "sort" has been inserted into RunGrepTest for the
3211 approprate test as a short-term fix. In the longer term there may be an
3212 alternative.
3214 3. I had an email from Eric Raymond about problems translating some of PCRE's
3215 man pages to HTML (despite the fact that I distribute HTML pages, some
3216 people do their own conversions for various reasons). The problems
3217 concerned the use of low-level troff macros .br and .in. I have therefore
3218 removed all such uses from the man pages (some were redundant, some could
3219 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3220 HTML has been updated to handle .nf/.fi and to complain if it encounters
3221 .br or .in.
3223 4. Updated comments in configure.ac that get placed in config.h.in and also
3224 arranged for config.h to be included in the distribution, with the name
3225 config.h.generic, for the benefit of those who have to compile without
3226 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3228 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3229 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3230 makevp.bat for the current PCRE, using the additional files
3231 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3233 6. A Windows user reported a minor discrepancy with test 2, which turned out
3234 to be caused by a trailing space on an input line that had got lost in his
3235 copy. The trailing space was an accident, so I've just removed it.
3237 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3238 that is needed.
3240 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3241 as "const" (a) because they are and (b) because it helps the PHP
3242 maintainers who have recently made a script to detect big data structures
3243 in the php code that should be moved to the .rodata section. I remembered
3244 to update Builducptable as well, so it won't revert if ucptable.h is ever
3245 re-created.
3247 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3248 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3249 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3250 support is not required. This saves 1.5-2K of code, which is important in
3251 some applications.
3253 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3254 so as not to refer to the tables, even though these functions will never be
3255 called when UTF-8 support is disabled. Otherwise there are problems with a
3256 shared library.
3258 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3260 (a) It was defining its arguments as char * instead of void *.
3262 (b) It was assuming that all moves were upwards in memory; this was true
3263 a long time ago when I wrote it, but is no longer the case.
3265 The emulated memove() is provided for those environments that have neither
3266 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3267 is clearly not the case, as these two bugs were recently reported.
3269 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3270 and Detrail to create the HTML documentation, the .txt form of the man
3271 pages, and it removes trailing spaces from listed files. It also creates
3272 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3273 case, it wraps all the #defines with #ifndefs. This script should be run
3274 before "make dist".
3276 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3277 with Unicode property support.
3279 (a) For a maximizing quantifier, if the two different cases of the
3280 character were of different lengths in their UTF-8 codings (there are
3281 some cases like this - I found 11), and the matching function had to
3282 back up over a mixture of the two cases, it incorrectly assumed they
3283 were both the same length.
3285 (b) When PCRE was configured to use the heap rather than the stack for
3286 recursion during matching, it was not correctly preserving the data for
3287 the other case of a UTF-8 character when checking ahead for a match
3288 while processing a minimizing repeat. If the check also involved
3289 matching a wide character, but failed, corruption could cause an
3290 erroneous result when trying to check for a repeat of the original
3291 character.
3293 13. Some tidying changes to the testing mechanism:
3295 (a) The RunTest script now detects the internal link size and whether there
3296 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3297 values substituted by "configure". (The RunGrepTest script already did
3298 this for UTF-8.) The configure.ac script no longer substitutes the
3299 relevant variables.
3301 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3302 with length and offset values. This means that the output is different
3303 for different internal link sizes. Test 2 is skipped for link sizes
3304 other than 2 because of this, bypassing the problem. Unfortunately,
3305 there was also a test in test 3 (the locale tests) that used /B and
3306 failed for link sizes other than 2. Rather than cut the whole test out,
3307 I have added a new /Z option to pcretest that replaces the length and
3308 offset values with spaces. This is now used to make test 3 independent
3309 of link size. (Test 2 will be tidied up later.)
3311 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3312 segmentation fault instead of returning the appropriate error message.
3314 15. In multiline mode when the newline sequence was set to "any", the pattern
3315 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3316 This doesn't seem right; it now treats the CRLF combination as the line
3317 ending, and so does not match in that case. It's only a pattern such as ^$
3318 that would hit this one: something like ^ABC$ would have failed after \r
3319 and then tried again after \r\n.
3321 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3322 in an attempt to make files that differ only in their line terminators
3323 compare equal. This works on Linux.
3325 17. Under certain error circumstances pcregrep might try to free random memory
3326 as it exited. This is now fixed, thanks to valgrind.
3328 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3329 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3330 was because its rules for how to advance for /g after matching an empty
3331 string at the end of a line did not allow for this case. They now check for
3332 it specially.
3334 20. pcretest is supposed to handle patterns and data of any length, by
3335 extending its buffers when necessary. It was getting this wrong when the
3336 buffer for a data line had to be extended.
3338 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3339 CRLF as a newline sequence.
3341 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3342 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3343 I have nevertheless tidied it up.
3345 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3347 24. Added a man page for pcre-config.
3350 Version 7.0 19-Dec-06
3351 ---------------------
3353 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3354 moving to gcc 4.1.1.
3356 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3357 sys/time.h, which is documented as needed for this function. It doesn't
3358 seem to matter on Linux, but it showed up on some releases of OS X.
3360 3. It seems that there are systems where bytes whose values are greater than
3361 127 match isprint() in the "C" locale. The "C" locale should be the
3362 default when a C program starts up. In most systems, only ASCII printing
3363 characters match isprint(). This difference caused the output from pcretest
3364 to vary, making some of the tests fail. I have changed pcretest so that:
3366 (a) When it is outputting text in the compiled version of a pattern, bytes
3367 other than 32-126 are always shown as hex escapes.
3369 (b) When it is outputting text that is a matched part of a subject string,
3370 it does the same, unless a different locale has been set for the match
3371 (using the /L modifier). In this case, it uses isprint() to decide.
3373 4. Fixed a major bug that caused incorrect computation of the amount of memory
3374 required for a compiled pattern when options that changed within the
3375 pattern affected the logic of the preliminary scan that determines the
3376 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3377 that the computed length was too small. The symptoms of this bug were
3378 either the PCRE error "internal error: code overflow" from pcre_compile(),
3379 or a glibc crash with a message such as "pcretest: free(): invalid next
3380 size (fast)". Examples of patterns that provoked this bug (shown in
3381 pcretest format) are:
3383 /(?-x: )/x
3384 /(?x)(?-x: \s*#\s*)/
3385 /((?i)[\x{c0}])/8
3386 /(?i:[\x{c0}])/8
3388 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3389 is now done differently.
3391 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3392 wrapper classes; (b) implement a new function in the C++ scanner that is
3393 more efficient than the old way of doing things because it avoids levels of
3394 recursion in the regex matching; (c) add a paragraph to the documentation
3395 for the FullMatch() function.
3397 6. The escape sequence \n was being treated as whatever was defined as
3398 "newline". Not only was this contrary to the documentation, which states
3399 that \n is character 10 (hex 0A), but it also went horribly wrong when
3400 "newline" was defined as CRLF. This has been fixed.
3402 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3403 was being set to -1 for the "end of line" case (supposedly a value that no
3404 character can have). Though this value is never used (the check for end of
3405 line is "zero bytes in current character"), it caused compiler complaints.
3406 I've changed it to 0xffffffff.
3408 8. In pcre_version.c, the version string was being built by a sequence of
3409 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3410 string (as it is for production releases) called a macro with an empty
3411 argument. The C standard says the result of this is undefined. The gcc
3412 compiler treats it as an empty string (which was what was wanted) but it is
3413 reported that Visual C gives an error. The source has been hacked around to
3414 avoid this problem.
3416 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3417 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3418 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3419 of them did).
3421 10. Originally, pcretest opened its input and output without "b"; then I was
3422 told that "b" was needed in some environments, so it was added for release
3423 5.0 to both the input and output. (It makes no difference on Unix-like
3424 systems.) Later I was told that it is wrong for the input on Windows. I've
3425 now abstracted the modes into two macros, to make it easier to fiddle with
3426 them, and removed "b" from the input mode under Windows.
3428 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3430 12. Added -help and --help to pcretest as an official way of being reminded
3431 of the options.
3433 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3434 and pcrecpp.cc because they annoy compilers at high warning levels.
3436 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3438 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3439 scripts are not C programs :-) and which was not noticed because it works
3440 on Linux.
3442 16. pcretest is supposed to handle any length of pattern and data line (as one
3443 line or as a continued sequence of lines) by extending its input buffer if
3444 necessary. This feature was broken for very long pattern lines, leading to
3445 a string of junk being passed to pcre_compile() if the pattern was longer
3446 than about 50K.
3448 17. I have done a major re-factoring of the way pcre_compile() computes the
3449 amount of memory needed for a compiled pattern. Previously, there was code
3450 that made a preliminary scan of the pattern in order to do this. That was
3451 OK when PCRE was new, but as the facilities have expanded, it has become
3452 harder and harder to keep it in step with the real compile phase, and there
3453 have been a number of bugs (see for example, 4 above). I have now found a
3454 cunning way of running the real compile function in a "fake" mode that
3455 enables it to compute how much memory it would need, while actually only
3456 ever using a few hundred bytes of working memory and without too many
3457 tests of the mode. This should make future maintenance and development
3458 easier. A side effect of this work is that the limit of 200 on the nesting
3459 depth of parentheses has been removed (though this was never a serious
3460 limitation, I suspect). However, there is a downside: pcre_compile() now
3461 runs more slowly than before (30% or more, depending on the pattern). I
3462 hope this isn't a big issue. There is no effect on runtime performance.
3464 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3465 newline (only possible for the last line of a file) and it was a
3466 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3468 19. Added additional timing features to pcretest. (1) The -tm option now times
3469 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3470 separate command line item, by a number that specifies the number of
3471 repeats to use when timing. The default is 50000; this gives better
3472 precision, but takes uncomfortably long for very large patterns.
3474 20. Extended pcre_study() to be more clever in cases where a branch of a
3475 subpattern has no definite first character. For example, (a*|b*)[cd] would
3476 previously give no result from pcre_study(). Now it recognizes that the
3477 first character must be a, b, c, or d.
3479 21. There was an incorrect error "recursive call could loop indefinitely" if
3480 a subpattern (or the entire pattern) that was being tested for matching an
3481 empty string contained only one non-empty item after a nested subpattern.
3482 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3483 incorrectly, because the \d was being skipped in the check.
3485 22. The pcretest program now has a new pattern option /B and a command line
3486 option -b, which is equivalent to adding /B to every pattern. This causes
3487 it to show the compiled bytecode, without the additional information that
3488 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3489 is the same as /B/I).
3491 23. A new optimization is now able automatically to treat some sequences such
3492 as a*b as a*+b. More specifically, if something simple (such as a character
3493 or a simple class like \d) has an unlimited quantifier, and is followed by
3494 something that cannot possibly match the quantified thing, the quantifier
3495 is automatically "possessified".
3497 24. A recursive reference to a subpattern whose number was greater than 39
3498 went wrong under certain circumstances in UTF-8 mode. This bug could also
3499 have affected the operation of pcre_study().
3501 25. Realized that a little bit of performance could be had by replacing
3502 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3504 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3506 27. Possessive quantifiers such as a++ were previously implemented by turning
3507 them into atomic groups such as ($>a+). Now they have their own opcodes,
3508 which improves performance. This includes the automatically created ones
3509 from 23 above.
3511 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3512 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3513 the first matched character to be a colon. This applied both to named and
3514 numbered groups.
3516 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3518 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3519 building PCRE on Windows easier, so I have included it in the distribution.
3521 31. There is now a check in pcretest against a ridiculously large number being
3522 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3523 loop, the loop is abandoned.
3525 32. Forward references to subpatterns in conditions such as (?(2)...) where
3526 subpattern 2 is defined later cause pcre_compile() to search forwards in
3527 the pattern for the relevant set of parentheses. This search went wrong
3528 when there were unescaped parentheses in a character class, parentheses
3529 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3531 33. "Subroutine" calls and backreferences were previously restricted to
3532 referencing subpatterns earlier in the regex. This restriction has now
3533 been removed.
3535 34. Added a number of extra features that are going to be in Perl 5.10. On the
3536 whole, these are just syntactic alternatives for features that PCRE had
3537 previously implemented using the Python syntax or my own invention. The
3538 other formats are all retained for compatibility.
3540 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3541 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3542 also .NET compatible.
3544 (b) A recursion or subroutine call to a named group can now be defined as
3545 (?&name) as well as (?P>name).
3547 (c) A backreference to a named group can now be defined as \k<name> or
3548 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3549 5.10, are also .NET compatible.
3551 (d) A conditional reference to a named group can now use the syntax
3552 (?(<name>) or (?('name') as well as (?(name).
3554 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3555 groups (named and numbered) that are never evaluated inline, but can be
3556 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3557 is always false. There may be only one alternative in such a group.
3559 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3560 as the simple (?(R). The condition is true only if the most recent
3561 recursion is that of the given number or name. It does not search out
3562 through the entire recursion stack.
3564 (g) The escape \gN or \g{N} has been added, where N is a positive or
3565 negative number, specifying an absolute or relative reference.
3567 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3568 some "unreachable code" warnings.
3570 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3571 things, this adds five new scripts.
3573 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3574 There were also incompatibilities regarding the handling of \Q..\E inside
3575 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3576 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3578 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3579 matches an empty string, and forcibly breaks the loop. There were bugs in
3580 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3581 against aaaa the result was just "a" rather than "aaaa", for example. Two
3582 separate and independent bugs (that affected different cases) have been
3583 fixed.
3585 39. Refactored the code to abolish the use of different opcodes for small
3586 capturing bracket numbers. This is a tidy that I avoided doing when I
3587 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3588 The new approach is not only tidier, it makes it possible to reduce the
3589 memory needed to fix the previous bug (38).
3591 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3592 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3593 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3594 mode.
3596 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3597 report.
3599 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3600 copy construction and assignment in the C++ wrapper.
3602 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3603 couple of bugs that could have given wrong results in the "--newline=crlf"
3604 case.
3606 44. Added a number of casts and did some reorganization of signed/unsigned int
3607 variables following suggestions from Dair Grant. Also renamed the variable
3608 "this" as "item" because it is a C++ keyword.
3610 45. Arranged for dftables to add
3612 #include "pcre_internal.h"
3614 to pcre_chartables.c because without it, gcc 4.x may remove the array
3615 definition from the final binary if PCRE is built into a static library and
3616 dead code stripping is activated.
3618 46. For an unanchored pattern, if a match attempt fails at the start of a
3619 newline sequence, and the newline setting is CRLF or ANY, and the next two
3620 characters are CRLF, advance by two characters instead of one.
3623 Version 6.7 04-Jul-06
3624 ---------------------
3626 1. In order to handle tests when input lines are enormously long, pcretest has
3627 been re-factored so that it automatically extends its buffers when
3628 necessary. The code is crude, but this _is_ just a test program. The
3629 default size has been increased from 32K to 50K.
3631 2. The code in pcre_study() was using the value of the re argument before
3632 testing it for NULL. (Of course, in any sensible call of the function, it
3633 won't be NULL.)
3635 3. The memmove() emulation function in pcre_internal.h, which is used on
3636 systems that lack both memmove() and bcopy() - that is, hardly ever -
3637 was missing a "static" storage class specifier.
3639 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3640 containing an extended class (one that cannot be represented by a bitmap
3641 because it contains high-valued characters or Unicode property items, e.g.
3642 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3643 pattern, but PCRE should not loop if you do not (it no longer does).
3644 [Detail: two cases were found: (a) a repeated subpattern containing an
3645 extended class; (b) a recursive reference to a subpattern that followed a
3646 previous extended class. It wasn't skipping over the extended class
3647 correctly when UTF-8 mode was not set.]
3649 5. A negated single-character class was not being recognized as fixed-length
3650 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3651 compile error "lookbehind assertion is not fixed length".
3653 6. The RunPerlTest auxiliary script was showing an unexpected difference
3654 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3655 write a Perl script that can interpret lines of an input file either as
3656 byte characters or as UTF-8, which is what "perltest" was being required to
3657 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3658 can't do is switch easily at run time between having the "use utf8;" pragma
3659 or not. In the end, I fudged it by using the RunPerlTest script to insert
3660 "use utf8;" explicitly for the UTF-8 tests.
3662 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3663 the end of the subject string, contrary to the documentation and to what
3664 Perl does. This was true of both matching functions. Now it matches only at
3665 the start of the subject and immediately after *internal* newlines.
3667 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3668 a pointer to an int instead of a pointer to an unsigned long int. This
3669 caused problems on 64-bit systems.
3671 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3672 instance of the 'standard' template library not being so standard".
3674 10. There was no check on the number of named subpatterns nor the maximum
3675 length of a subpattern name. The product of these values is used to compute
3676 the size of the memory block for a compiled pattern. By supplying a very
3677 long subpattern name and a large number of named subpatterns, the size
3678 computation could be caused to overflow. This is now prevented by limiting
3679 the length of names to 32 characters, and the number of named subpatterns
3680 to 10,000.
3682 11. Subpatterns that are repeated with specific counts have to be replicated in
3683 the compiled pattern. The size of memory for this was computed from the
3684 length of the subpattern and the repeat count. The latter is limited to
3685 65535, but there was no limit on the former, meaning that integer overflow
3686 could in principle occur. The compiled length of a repeated subpattern is
3687 now limited to 30,000 bytes in order to prevent this.
3689 12. Added the optional facility to have named substrings with the same name.
3691 13. Added the ability to use a named substring as a condition, using the
3692 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3693 are numbers (not recommended). Forward references are permitted.
3695 14. Added forward references in named backreferences (if you see what I mean).
3697 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3698 pattern could run off the end of the subject. For example, the pattern
3699 "(?s)(.{1,5})"8 did this with the subject "ab".
3701 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3702 PCRE_CASELESS was set when matching characters that were quantified with ?
3703 or *.
3705 17. A character class other than a single negated character that had a minimum
3706 but no maximum quantifier - for example [ab]{6,} - was not handled
3707 correctly by pce_dfa_exec(). It would match only one character.
3709 18. A valid (though odd) pattern that looked like a POSIX character
3710 class but used an invalid character after [ (for example [[,abc,]]) caused
3711 pcre_compile() to give the error "Failed: internal error: code overflow" or
3712 in some cases to crash with a glibc free() error. This could even happen if
3713 the pattern terminated after [[ but there just happened to be a sequence of
3714 letters, a binary zero, and a closing ] in the memory that followed.
3716 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3717 over the years. Originally (before any Unicode support), just the bottom 8
3718 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3719 output from "man perlunicode" includes this:
3721 The regular expression compiler produces polymorphic opcodes. That
3722 is, the pattern adapts to the data and automatically switches to
3723 the Unicode character scheme when presented with Unicode data--or
3724 instead uses a traditional byte scheme when presented with byte
3725 data.
3727 Sadly, a wide octal escape does not cause a switch, and in a string with
3728 no other multibyte characters, these octal escapes are treated as before.
3729 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3730 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3731 Unicode string.
3733 I have not perpetrated such confusion in PCRE. Up till now, it took just
3734 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3735 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3736 translate to the appropriate multibyte character.
3738 29. Applied some refactoring to reduce the number of warnings from Microsoft
3739 and Borland compilers. This has included removing the fudge introduced
3740 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3741 a warning about an unused variable.
3743 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3744 characters since release 4.0, because Perl (from release 5.004) does not.
3745 [Or at least, is documented not to: some releases seem to be in conflict
3746 with the documentation.] However, when a pattern was studied with
3747 pcre_study() and all its branches started with \s, PCRE still included VT
3748 as a possible starting character. Of course, this did no harm; it just
3749 caused an unnecessary match attempt.
3751 22. Removed a now-redundant internal flag bit that recorded the fact that case
3752 dependency changed within the pattern. This was once needed for "required
3753 byte" processing, but is no longer used. This recovers a now-scarce options
3754 bit. Also moved the least significant internal flag bit to the most-
3755 significant bit of the word, which was not previously used (hangover from
3756 the days when it was an int rather than a uint) to free up another bit for
3757 the future.
3759 23. Added support for CRLF line endings as well as CR and LF. As well as the
3760 default being selectable at build time, it can now be changed at runtime
3761 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3762 specify that it is scanning data with non-default line endings.
3764 24. Changed the definition of CXXLINK to make it agree with the definition of
3765 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3767 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3768 recursions. This makes a big different to stack usage for some patterns.
3770 26. If a subpattern containing a named recursion or subroutine reference such
3771 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3772 the space required for the compiled pattern went wrong and gave too small a
3773 value. Depending on the environment, this could lead to "Failed: internal
3774 error: code overflow at offset 49" or "glibc detected double free or
3775 corruption" errors.
3777 27. Applied patches from Google (a) to support the new newline modes and (b) to
3778 advance over multibyte UTF-8 characters in GlobalReplace.
3780 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3781 difference for some implementation of PCRE in some Windows version.
3783 29. Added some extra testing facilities to pcretest:
3785 \q<number> in a data line sets the "match limit" value
3786 \Q<number> in a data line sets the "match recursion limt" value
3787 -S <number> sets the stack size, where <number> is in megabytes
3789 The -S option isn't available for Windows.
3792 Version 6.6 06-Feb-06
3793 ---------------------
3795 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3796 in pcreposix.h. I have copied the definition from pcre.h.
3798 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3799 because pcre.h is no longer a built file.
3801 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3802 not normally included in the compiled code.
3805 Version 6.5 01-Feb-06
3806 ---------------------
3808 1. When using the partial match feature with pcre_dfa_exec(), it was not
3809 anchoring the second and subsequent partial matches at the new starting
3810 point. This could lead to incorrect results. For example, with the pattern
3811 /1234/, partially matching against "123" and then "a4" gave a match.
3813 2. Changes to pcregrep:
3815 (a) All non-match returns from pcre_exec() were being treated as failures
3816 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3817 error message is output. Some extra information is given for the
3819 probably the only errors that are likely to be caused by users (by
3820 specifying a regex that has nested indefinite repeats, for instance).
3821 If there are more than 20 of these errors, pcregrep is abandoned.
3823 (b) A binary zero was treated as data while matching, but terminated the
3824 output line if it was written out. This has been fixed: binary zeroes
3825 are now no different to any other data bytes.
3827 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3828 used to set a locale for matching. The --locale=xxxx long option has
3829 been added (no short equivalent) to specify a locale explicitly on the
3830 pcregrep command, overriding the environment variables.
3832 (d) When -B was used with -n, some line numbers in the output were one less
3833 than they should have been.
3835 (e) Added the -o (--only-matching) option.
3837 (f) If -A or -C was used with -c (count only), some lines of context were
3838 accidentally printed for the final match.
3840 (g) Added the -H (--with-filename) option.
3842 (h) The combination of options -rh failed to suppress file names for files
3843 that were found from directory arguments.
3845 (i) Added the -D (--devices) and -d (--directories) options.
3847 (j) Added the -F (--fixed-strings) option.
3849 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3851 (l) Added the --colo(u)r option.
3853 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3854 is not present by default.
3856 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3857 items such as (?R) or (?1), when the recursion could match a number of
3858 alternatives. If it matched one of the alternatives, but subsequently,
3859 outside the recursion, there was a failure, the code tried to back up into
3860 the recursion. However, because of the way PCRE is implemented, this is not
3861 possible, and the result was an incorrect result from the match.
3863 In order to prevent this happening, the specification of recursion has
3864 been changed so that all such subpatterns are automatically treated as
3865 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3867 4. I had overlooked the fact that, in some locales, there are characters for
3868 which isalpha() is true but neither isupper() nor islower() are true. In
3869 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3870 and ordfeminine) are like this. This affected the treatment of \w and \W
3871 when they appeared in character classes, but not when they appeared outside
3872 a character class. The bit map for "word" characters is now created
3873 separately from the results of isalnum() instead of just taking it from the
3874 upper, lower, and digit maps. (Plus the underscore character, of course.)
3876 5. The above bug also affected the handling of POSIX character classes such as
3877 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3878 permanent tables. Instead, the bit maps for such a class were previously
3879 created as the appropriate unions of the upper, lower, and digit bitmaps.
3880 Now they are created by subtraction from the [[:word:]] class, which has
3881 its own bitmap.
3883 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3884 It is created by subtracting the vertical space characters (\x09, \x0a,
3885 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3886 subtraction was done in the overall bitmap for a character class, meaning
3887 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3888 be recognized. This bug has been fixed.
3890 7. Patches from the folks at Google:
3892 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3893 real life, but is still worth protecting against".
3895 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3896 regular expressions".
3898 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3899 have it.
3901 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3902 "configure" and the latter not, in order to fix a problem somebody had
3903 with compiling the Arg class on HP-UX.
3905 (e) Improve the error-handling of the C++ wrapper a little bit.
3907 (f) New tests for checking recursion limiting.
3909 8. The pcre_memmove() function, which is used only if the environment does not
3910 have a standard memmove() function (and is therefore rarely compiled),
3911 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3912 returning a result (though PCRE never actually uses the result).
3914 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3915 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3916 returned instead of calling malloc() with an overflowing number that would
3917 most likely cause subsequent chaos.
3919 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3921 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3922 with this option is matched, the nmatch and pmatch options of regexec() are
3923 ignored.
3925 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3926 provided in case anyone wants to the the POSIX interface with UTF-8
3927 strings.
3929 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3930 C++ linking (needed for some HP-UX environments).
3932 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3933 (unused parameter) and in the pcre_printint() function (omitted "default"
3934 switch label when the default is to do nothing).
3936 15. Added some code to make it possible, when PCRE is compiled as a C++
3937 library, to replace subject pointers for pcre_exec() with a smart pointer
3938 class, thus making it possible to process discontinuous strings.
3940 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3941 much the same function. They were added by different people who were trying
3942 to make PCRE easy to compile on non-Unix systems. It has been suggested
3943 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3944 for compiling on Windows systems. I have therefore replaced it with
3945 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3946 defaults to "extern" for C or "extern C" for C++, which works fine on
3947 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3948 SCOPE with something explicit in config.h. In addition:
3950 (a) pcreposix.h still had just "extern" instead of either of these macros;
3951 I have replaced it with PCRE_DATA_SCOPE.
3953 (b) Functions such as _pcre_xclass(), which are internal to the library,
3954 but external in the C sense, all had PCRE_EXPORT in their definitions.
3955 This is apparently wrong for the Windows case, so I have removed it.
3956 (It makes no difference on Unix-like systems.)
3958 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3959 of recursive calls to match(). This is different to MATCH_LIMIT because
3960 that limits the total number of calls to match(), not all of which increase
3961 the depth of recursion. Limiting the recursion depth limits the amount of
3962 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3963 when PCRE is compiled, and changed at run time. A patch from Google adds
3964 this functionality to the C++ interface.
3966 18. Changes to the handling of Unicode character properties:
3968 (a) Updated the table to Unicode 4.1.0.
3970 (b) Recognize characters that are not in the table as "Cn" (undefined).
3972 (c) I revised the way the table is implemented to a much improved format
3973 which includes recognition of ranges. It now supports the ranges that
3974 are defined in UnicodeData.txt, and it also amalgamates other
3975 characters into ranges. This has reduced the number of entries in the
3976 table from around 16,000 to around 3,000, thus reducing its size
3977 considerably. I realized I did not need to use a tree structure after
3978 all - a binary chop search is just as efficient. Having reduced the
3979 number of entries, I extended their size from 6 bytes to 8 bytes to
3980 allow for more data.
3982 (d) Added support for Unicode script names via properties such as \p{Han}.
3984 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3985 matching that character.
3987 20. When matching a repeated Unicode property with a minimum greater than zero,
3988 (for example \pL{2,}), PCRE could look past the end of the subject if it
3989 reached it while seeking the minimum number of characters. This could
3990 happen only if some of the characters were more than one byte long, because
3991 there is a check for at least the minimum number of bytes.
3993 21. Refactored the implementation of \p and \P so as to be more general, to
3994 allow for more different types of property in future. This has changed the
3995 compiled form incompatibly. Anybody with saved compiled patterns that use
3996 \p or \P will have to recompile them.
3998 22. Added "Any" and "L&" to the supported property types.
4000 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4001 but give a compile time error if the value is greater than 0xff.
4003 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4004 accidentally not being installed or uninstalled.
4006 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4007 made were to insert the current release number. This seemed silly, because
4008 it made things harder for people building PCRE on systems that don't run
4009 "configure". I have turned pcre.h into a distributed file, no longer built
4010 by "configure", with the version identification directly included. There is
4011 no longer a pcre.h.in file.
4013 However, this change necessitated a change to the pcre-config script as
4014 well. It is built from pcre-config.in, and one of the substitutions was the
4015 release number. I have updated configure.ac so that ./configure now finds
4016 the release number by grepping pcre.h.
4018 26. Added the ability to run the tests under valgrind.
4021 Version 6.4 05-Sep-05
4022 ---------------------
4024 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4025 "--" to be printed when multiple files were scanned, even when none of the
4026 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4027 consider it to be a bug, and have restored the previous behaviour.
4029 2. A couple of code tidies to get rid of compiler warnings.
4031 3. The pcretest program used to cheat by referring to symbols in the library
4032 whose names begin with _pcre_. These are internal symbols that are not
4033 really supposed to be visible externally, and in some environments it is
4034 possible to suppress them. The cheating is now confined to including
4035 certain files from the library's source, which is a bit cleaner.
4037 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4038 file's purpose clearer.
4040 5. Reorganized pcre_ucp_findchar().
4043 Version 6.3 15-Aug-05
4044 ---------------------
4046 1. The file libpcre.pc.in did not have general read permission in the tarball.
4048 2. There were some problems when building without C++ support:
4050 (a) If C++ support was not built, "make install" and "make test" still
4051 tried to test it.
4053 (b) There were problems when the value of CXX was explicitly set. Some
4054 changes have been made to try to fix these, and ...
4056 (c) --disable-cpp can now be used to explicitly disable C++ support.
4058 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4059 backslash in a target when C++ was disabled. This confuses some
4060 versions of "make", apparently. Using an intermediate variable solves
4061 this. (Same for CPP_LOBJ.)
4063 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4064 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4065 necessary on certain architectures.
4067 4. Added a setting of -export-symbols-regex to the link command to remove
4068 those symbols that are exported in the C sense, but actually are local
4069 within the library, and not documented. Their names all begin with
4070 "_pcre_". This is not a perfect job, because (a) we have to except some
4071 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4072 available (and never for static libraries). I have made a note to try to
4073 find a way round (a) in the future.
4076 Version 6.2 01-Aug-05
4077 ---------------------
4079 1. There was no test for integer overflow of quantifier values. A construction
4080 such as {1111111111111111} would give undefined results. What is worse, if
4081 a minimum quantifier for a parenthesized subpattern overflowed and became
4082 negative, the calculation of the memory size went wrong. This could have
4083 led to memory overwriting.
4085 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4087 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4088 operating environments where this matters.
4090 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4091 PCRE options from within the C++ wrapper.
4093 5. Named capturing subpatterns were not being correctly counted when a pattern
4094 was compiled. This caused two problems: (a) If there were more than 100
4095 such subpatterns, the calculation of the memory needed for the whole
4096 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4097 back references of the form \12, where the number was greater than 9, were
4098 not recognized as back references, even though there were sufficient
4099 previous subpatterns.
4101 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4102 versions of gcc, e.g. 2.95.4.
4105 Version 6.1 21-Jun-05
4106 ---------------------
4108 1. There was one reference to the variable "posix" in pcretest.c that was not
4109 surrounded by "#if !defined NOPOSIX".
4111 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4112 the cross-check on the old pcre_info() function, for the benefit of the
4113 cut-down version of PCRE that is currently imported into Exim.
4115 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4116 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4117 patterns that start (?i) but I don't think that matters. The use of (?i) is
4118 just an example; this all applies to the other options as well.
4120 4. Since libtool seems to echo the compile commands it is issuing, the output
4121 from "make" can be reduced a bit by putting "@" in front of each libtool
4122 compile command.
4124 5. Patch from the folks at Google for configure.in to be a bit more thorough
4125 in checking for a suitable C++ installation before trying to compile the
4126 C++ stuff. This should fix a reported problem when a compiler was present,
4127 but no suitable headers.
4129 6. The man pages all had just "PCRE" as their title. I have changed them to
4130 be the relevant file name. I have also arranged that these names are
4131 retained in the file doc/pcre.txt, which is a concatenation in text format
4132 of all the man pages except the little individual ones for each function.
4134 7. The NON-UNIX-USE file had not been updated for the different set of source
4135 files that come with release 6. I also added a few comments about the C++
4136 wrapper.
4139 Version 6.0 07-Jun-05
4140 ---------------------
4142 1. Some minor internal re-organization to help with my DFA experiments.
4144 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4145 didn't matter for the library itself when fully configured, but did matter
4146 when compiling without UCP support, or within Exim, where the ucp files are
4147 not imported.
4149 3. Refactoring of the library code to split up the various functions into
4150 different source modules. The addition of the new DFA matching code (see
4151 below) to a single monolithic source would have made it really too
4152 unwieldy, quite apart from causing all the code to be include in a
4153 statically linked application, when only some functions are used. This is
4154 relevant even without the DFA addition now that patterns can be compiled in
4155 one application and matched in another.
4157 The downside of splitting up is that there have to be some external
4158 functions and data tables that are used internally in different modules of
4159 the library but which are not part of the API. These have all had their
4160 names changed to start with "_pcre_" so that they are unlikely to clash
4161 with other external names.
4163 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4164 a different (DFA) algorithm. Although it is slower than the original
4165 function, it does have some advantages for certain types of matching
4166 problem.
4168 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4169 including restarting after a partial match.
4171 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4172 defined when compiling for Windows was sent to me. I have put it into the
4173 code, though I have no means of testing or verifying it.
4175 7. Added the pcre_refcount() auxiliary function.
4177 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4178 match before or at the first newline in the subject string. In pcretest,
4179 the /f option on a pattern can be used to set this.
4181 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4182 would behave wrongly. This has been present in PCRE since release 4.0.
4184 10. A number of changes to the pcregrep command:
4186 (a) Refactored how -x works; insert ^(...)$ instead of setting
4187 PCRE_ANCHORED and checking the length, in preparation for adding
4188 something similar for -w.
4190 (b) Added the -w (match as a word) option.
4192 (c) Refactored the way lines are read and buffered so as to have more
4193 than one at a time available.
4195 (d) Implemented a pcregrep test script.
4197 (e) Added the -M (multiline match) option. This allows patterns to match
4198 over several lines of the subject. The buffering ensures that at least
4199 8K, or the rest of the document (whichever is the shorter) is available
4200 for matching (and similarly the previous 8K for lookbehind assertions).
4202 (f) Changed the --help output so that it now says
4204 -w, --word-regex(p)
4206 instead of two lines, one with "regex" and the other with "regexp"
4207 because that confused at least one person since the short forms are the
4208 same. (This required a bit of code, as the output is generated
4209 automatically from a table. It wasn't just a text change.)
4211 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4212 option but starts with a hyphen. Could be a pattern or a path name
4213 starting with a hyphen, for instance.
4215 (h) "-" can be given as a file name to represent stdin.
4217 (i) When file names are being printed, "(standard input)" is used for
4218 the standard input, for compatibility with GNU grep. Previously
4219 "<stdin>" was used.
4221 (j) The option --label=xxx can be used to supply a name to be used for
4222 stdin when file names are being printed. There is no short form.
4224 (k) Re-factored the options decoding logic because we are going to add
4225 two more options that take data. Such options can now be given in four
4226 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4228 (l) Added the -A, -B, and -C options for requesting that lines of context
4229 around matches be printed.
4231 (m) Added the -L option to print the names of files that do not contain
4232 any matching lines, that is, the complement of -l.
4234 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4235 continue to scan other files.
4237 (o) The -s option was incorrectly implemented. For compatibility with other
4238 greps, it now suppresses the error message for a non-existent or non-
4239 accessible file (but not the return code). There is a new option called
4240 -q that suppresses the output of matching lines, which was what -s was
4241 previously doing.
4243 (p) Added --include and --exclude options to specify files for inclusion
4244 and exclusion when recursing.
4246 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4247 Hopefully, it now does.
4249 12. Missing cast in pcre_study().
4251 13. Added an "uninstall" target to the makefile.
4253 14. Replaced "extern" in the function prototypes in Makefile.in with
4254 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4255 world, but is set differently for Windows.
4257 15. Added a second compiling function called pcre_compile2(). The only
4258 difference is that it has an extra argument, which is a pointer to an
4259 integer error code. When there is a compile-time failure, this is set
4260 non-zero, in addition to the error test pointer being set to point to an
4261 error message. The new argument may be NULL if no error number is required
4262 (but then you may as well call pcre_compile(), which is now just a
4263 wrapper). This facility is provided because some applications need a
4264 numeric error indication, but it has also enabled me to tidy up the way
4265 compile-time errors are handled in the POSIX wrapper.
4267 16. Added VPATH=.libs to the makefile; this should help when building with one
4268 prefix path and installing with another. (Or so I'm told by someone who
4269 knows more about this stuff than I do.)
4271 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4272 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4273 match everything, including newlines. This is not POSIX-compatible, but
4274 somebody wanted the feature. From pcretest it can be activated by using
4275 both the P and the s flags.
4277 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4279 19. libpcre.pc was being incorrectly installed as executable.
4281 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4282 it now also looks for '\r' so that it will work unmodified on Windows.
4284 21. Added Google's contributed C++ wrapper to the distribution.
4286 22. Added some untidy missing memory free() calls in pcretest, to keep
4287 Electric Fence happy when testing.
4291 Version 5.0 13-Sep-04
4292 ---------------------
4294 1. Internal change: literal characters are no longer packed up into items
4295 containing multiple characters in a single byte-string. Each character
4296 is now matched using a separate opcode. However, there may be more than one
4297 byte in the character in UTF-8 mode.
4299 2. The pcre_callout_block structure has two new fields: pattern_position and
4300 next_item_length. These contain the offset in the pattern to the next match
4301 item, and its length, respectively.
4303 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4304 insertion of callouts before each pattern item. Added the /C option to
4305 pcretest to make use of this.
4307 4. On the advice of a Windows user, the lines
4309 #if defined(_WIN32) || defined(WIN32)
4310 _setmode( _fileno( stdout ), 0x8000 );
4311 #endif /* defined(_WIN32) || defined(WIN32) */
4313 have been added to the source of pcretest. This apparently does useful
4314 magic in relation to line terminators.
4316 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4317 for the benefit of those environments where the "b" makes a difference.
4319 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4320 to know about it. I have put a hack into configure.in that adds in code
4321 to set GCC=yes if CC=icc. This seems to end up at a point in the
4322 generated configure script that is early enough to affect the setting of
4323 compiler options, which is what is needed, but I have no means of testing
4324 whether it really works. (The user who reported this had patched the
4325 generated configure script, which of course I cannot do.)
4327 LATER: After change 22 below (new libtool files), the configure script
4328 seems to know about icc (and also ecc). Therefore, I have commented out
4329 this hack in configure.in.
4331 7. Added support for pkg-config (2 patches were sent in).
4333 8. Negated POSIX character classes that used a combination of internal tables
4334 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4335 [[:^ascii]]. Typically, they would match almost any characters. The other
4336 POSIX classes were not broken in this way.
4338 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4339 to find the match, as PCRE was deluded into thinking that the match had to
4340 start at the start point or following a newline. The same bug applied to
4341 patterns with negative forward assertions or any backward assertions
4342 preceding ".*" at the start, unless the pattern required a fixed first
4343 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4345 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4346 starting at the last subject character, bytes beyond the end of the subject
4347 string were read.
4349 11. Renamed the variable "class" as "classbits" to make life easier for C++
4350 users. (Previously there was a macro definition, but it apparently wasn't
4351 enough.)
4353 12. Added the new field "tables" to the extra data so that tables can be passed
4354 in at exec time, or the internal tables can be re-selected. This allows
4355 a compiled regex to be saved and re-used at a later time by a different
4356 program that might have everything at different addresses.
4358 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4359 -R library as well as a -L library.
4361 14. The debugging options of pcretest (-d on the command line or D on a
4362 pattern) showed incorrect output for anything following an extended class
4363 that contained multibyte characters and which was followed by a quantifier.
4365 15. Added optional support for general category Unicode character properties
4366 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4367 support. It adds about 90K to the size of the library. The meanings of the
4368 inbuilt class escapes such as \d and \s have NOT been changed.
4370 16. Updated pcredemo.c to include calls to free() to release the memory for the
4371 compiled pattern.
4373 17. The generated file chartables.c was being created in the source directory
4374 instead of in the building directory. This caused the build to fail if the
4375 source directory was different from the building directory, and was
4376 read-only.
4378 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4379 file. No doubt somebody will tell me if they don't make sense... Also added
4380 Dan Mooney's comments about building on OpenVMS.
4382 19. Added support for partial matching via the PCRE_PARTIAL option for
4383 pcre_exec() and the \P data escape in pcretest.
4385 20. Extended pcretest with 3 new pattern features:
4387 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4388 write the compiled pattern to the file whose name is "rest-of-line".
4389 This is a straight binary dump of the data, with the saved pointer to
4390 the character tables forced to be NULL. The study data, if any, is
4391 written too. After writing, pcretest reads a new pattern.
4393 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4394 compiled pattern from the given file. There must not be any
4395 occurrences of "<" in the file name (pretty unlikely); if there are,
4396 pcretest will instead treat the initial "<" as a pattern delimiter.
4397 After reading in the pattern, pcretest goes on to read data lines as
4398 usual.
4400 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4401 and 16-bit fields in a compiled pattern, to simulate a pattern that
4402 was compiled on a host of opposite endianness.
4404 21. The pcre-exec() function can now cope with patterns that were compiled on
4405 hosts of opposite endianness, with this restriction:
4407 As for any compiled expression that is saved and used later, the tables
4408 pointer field cannot be preserved; the extra_data field in the arguments
4409 to pcre_exec() should be used to pass in a tables address if a value
4410 other than the default internal tables were used at compile time.
4412 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4413 now diagnosed as an error. Previously, most of the time, a negative number
4414 would have been treated as zero, but if in addition "ovector" was passed as
4415 NULL, a crash could occur.
4417 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4418 new versions from the libtool 1.5 distribution (the last one is a copy of
4419 a file called libtool.m4). This seems to have fixed the need to patch
4420 "configure" to support Darwin 1.3 (which I used to do). However, I still
4421 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4422 workstation).
4424 24. Changed the PCRE licence to be the more standard "BSD" licence.
4427 Version 4.5 01-Dec-03
4428 ---------------------
4430 1. There has been some re-arrangement of the code for the match() function so
4431 that it can be compiled in a version that does not call itself recursively.
4432 Instead, it keeps those local variables that need separate instances for
4433 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4434 needs to "recurse". Keeping track of where control must go is done by means
4435 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4436 hide most of the details from the main code, and operates only if
4437 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4438 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4439 operating.
4441 To make it easier for callers to provide specially tailored get/free
4442 functions for this usage, two new functions, pcre_stack_malloc, and
4443 pcre_stack_free, are used. They are always called in strict stacking order,
4444 and the size of block requested is always the same.
4446 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4447 PCRE has been compiled to use the stack or the heap for recursion. The
4448 -C option of pcretest uses this to show which version is compiled.
4450 A new data escape \S, is added to pcretest; it causes the amounts of store
4451 obtained and freed by both kinds of malloc/free at match time to be added
4452 to the output.
4454 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4455 what's available on my current Linux desktop machine.
4457 3. When matching a UTF-8 string, the test for a valid string at the start has
4458 been extended. If start_offset is not zero, PCRE now checks that it points
4459 to a byte that is the start of a UTF-8 character. If not, it returns
4460 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4461 this is necessary because there may be backward assertions in the pattern.
4462 When matching the same subject several times, it may save resources to use
4463 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4465 4. The code for checking the validity of UTF-8 strings has been tightened so
4466 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4467 containing "overlong sequences".
4469 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4470 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4471 should have been "&", but it just so happened that all the cases this let
4472 through by mistake were picked up later in the function.
4474 6. I had used a variable called "isblank" - this is a C99 function, causing
4475 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4477 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4478 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4479 so the version is shown; (c) stop "make test" echoing "./RunTest".
4481 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4483 9. The prototype for memmove() for systems that don't have it was using
4484 size_t, but the inclusion of the header that defines size_t was later. I've
4485 moved the #includes for the C headers earlier to avoid this.
4487 10. Added some adjustments to the code to make it easier to compiler on certain
4488 special systems:
4490 (a) Some "const" qualifiers were missing.
4491 (b) Added the macro EXPORT before all exported functions; by default this
4492 is defined to be empty.
4493 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4494 that it reads its output file name as an argument instead of writing
4495 to the standard output and assuming this can be redirected.
4497 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4498 class containing characters with values greater than 255, PCRE compilation
4499 went into a loop.
4501 12. A recursive reference to a subpattern that was within another subpattern
4502 that had a minimum quantifier of zero caused PCRE to crash. For example,
4503 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4504 recursion. If the recursively-called subpattern itself had a zero repeat,
4505 that was OK.
4507 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4508 buffer into which it was copied (for escape processing) was still set at
4509 1024, so long lines caused crashes.
4511 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4512 "internal error: code overflow...". This applied to any character class
4513 that was followed by a possessive quantifier.
4515 15. Modified the Makefile to add libpcre.la as a prerequisite for
4516 libpcreposix.la because I was told this is needed for a parallel build to
4517 work.
4519 16. If a pattern that contained .* following optional items at the start was
4520 studied, the wrong optimizing data was generated, leading to matching
4521 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4522 matching string must start with a or b or c. The correct conclusion for
4523 this pattern is that a match can start with any character.
4526 Version 4.4 13-Aug-03
4527 ---------------------
4529 1. In UTF-8 mode, a character class containing characters with values between
4530 127 and 255 was not handled correctly if the compiled pattern was studied.
4531 In fixing this, I have also improved the studying algorithm for such
4532 classes (slightly).
4534 2. Three internal functions had redundant arguments passed to them. Removal
4535 might give a very teeny performance improvement.
4537 3. Documentation bug: the value of the capture_top field in a callout is *one
4538 more than* the number of the hightest numbered captured substring.
4540 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4541 in incorrectly linking with a previously installed version. They now link
4542 explicitly with libpcre.la.
4544 5. configure.in no longer needs to recognize Cygwin specially.
4546 6. A problem in pcre.in for Windows platforms is fixed.
4548 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4549 pcretest, it used to include the size of the study block as part of its
4550 output. Unfortunately, the structure contains a field that has a different
4551 size on different hardware architectures. This meant that the tests that
4552 showed this size failed. As the block is currently always of a fixed size,
4553 this information isn't actually particularly useful in pcretest output, so
4554 I have just removed it.
4556 8. Three pre-processor statements accidentally did not start in column 1.
4557 Sadly, there are *still* compilers around that complain, even though
4558 standard C has not required this for well over a decade. Sigh.
4560 9. In pcretest, the code for checking callouts passed small integers in the
4561 callout_data field, which is a void * field. However, some picky compilers
4562 complained about the casts involved for this on 64-bit systems. Now
4563 pcretest passes the address of the small integer instead, which should get
4564 rid of the warnings.
4566 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4567 both compile and run time, and gives an error if an invalid UTF-8 sequence
4568 is found. There is a option for disabling this check in cases where the
4569 string is known to be correct and/or the maximum performance is wanted.
4571 11. In response to a bug report, I changed one line in Makefile.in from
4573 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4574 to
4575 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4577 to look similar to other lines, but I have no way of telling whether this
4578 is the right thing to do, as I do not use Windows. No doubt I'll get told
4579 if it's wrong...
4582 Version 4.3 21-May-03
4583 ---------------------
4585 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4586 Makefile.
4588 2. Some refactoring to improve the quality of the code:
4590 (i) The utf8_table... variables are now declared "const".
4592 (ii) The code for \cx, which used the "case flipping" table to upper case
4593 lower case letters, now just substracts 32. This is ASCII-specific,
4594 but the whole concept of \cx is ASCII-specific, so it seems
4595 reasonable.
4597 (iii) PCRE was using its character types table to recognize decimal and
4598 hexadecimal digits in the pattern. This is silly, because it handles
4599 only 0-9, a-f, and A-F, but the character types table is locale-
4600 specific, which means strange things might happen. A private
4601 table is now used for this - though it costs 256 bytes, a table is
4602 much faster than multiple explicit tests. Of course, the standard
4603 character types table is still used for matching digits in subject
4604 strings against \d.
4606 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4607 ending in _t are). So I've renamed it as ESC_tee.
4609 3. The first argument for regexec() in the POSIX wrapper should have been
4610 defined as "const".
4612 4. Changed pcretest to use malloc() for its buffers so that they can be
4613 Electric Fenced for debugging.
4615 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4616 to read one or more bytes before the start of the subject string. Often this
4617 had no effect on PCRE's behaviour, but in some circumstances it could
4618 provoke a segmentation fault.
4620 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4621 to try to read one or more bytes before the start of the subject string.
4623 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4624 UTF-8 support could misbehave in various ways if the subject string
4625 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4626 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4627 back over UTF-8 characters.)
4630 Version 4.2 14-Apr-03
4631 ---------------------
4633 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4635 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4636 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4637 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4638 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4639 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4641 Note: automatic setting of the BUILD variables is not yet working
4642 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4643 compile-time but not at link-time
4644 [LINK]: use for linking executables only
4645 make different versions for Windows and non-Windows
4646 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4647 libraries
4648 [LINK_FOR_BUILD]: new variable
4649 [OBJEXT]: use throughout
4650 [EXEEXT]: use throughout
4651 <winshared>: new target
4652 <wininstall>: new target
4653 <dftables.o>: use native compiler
4654 <dftables>: use native linker
4655 <install>: handle Windows platform correctly
4656 <clean>: ditto
4657 <check>: ditto
4658 copy DLL to top builddir before testing
4660 As part of these changes, -no-undefined was removed again. This was reported
4661 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4662 in any case.
4664 3. Some tidies to get rid of compiler warnings:
4666 . In the match_data structure, match_limit was an unsigned long int, whereas
4667 match_call_count was an int. I've made them both unsigned long ints.
4669 . In pcretest the fact that a const uschar * doesn't automatically cast to
4670 a void * provoked a warning.
4672 . Turning on some more compiler warnings threw up some "shadow" variables
4673 and a few more missing casts.
4675 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4676 option, a class that contained a single character with a value between 128
4677 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4679 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4680 option, a class that contained several characters, but with at least one
4681 whose value was between 128 and 255 caused PCRE to crash.
4684 Version 4.1 12-Mar-03
4685 ---------------------
4687 1. Compiling with gcc -pedantic found a couple of places where casts were
4688 needed, and a string in dftables.c that was longer than standard compilers are
4689 required to support.
4691 2. Compiling with Sun's compiler found a few more places where the code could
4692 be tidied up in order to avoid warnings.
4694 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4695 first of these names is deprecated in the latest Autoconf in favour of the name
4696 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4697 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4698 analogy I have changed it to CFLAGS_FOR_BUILD.
4700 4. Added -no-undefined to the linking command in the Makefile, because this is
4701 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4702 linking step for the pcreposix library.
4704 5. PCRE was failing to diagnose the case of two named groups with the same
4705 name.
4707 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4708 literal character that is needed in the subject for a match, and scans along to
4709 ensure that it is present before embarking on the full matching process. This
4710 saves time in cases of nested unlimited repeats that are never going to match.
4711 Problem: the scan can take a lot of time if the subject is very long (e.g.
4712 megabytes), thus penalizing straightforward matches. It is now done only if the
4713 amount of subject to be scanned is less than 1000 bytes.
4715 7. A lesser problem with the same optimization is that it was recording the
4716 first character of an anchored pattern as "needed", thus provoking a search
4717 right along the subject, even when the first match of the pattern was going to
4718 fail. The "needed" character is now not set for anchored patterns, unless it
4719 follows something in the pattern that is of non-fixed length. Thus, it still
4720 fulfils its original purpose of finding quick non-matches in cases of nested
4721 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4724 Version 4.0 17-Feb-03
4725 ---------------------
4727 1. If a comment in an extended regex that started immediately after a meta-item
4728 extended to the end of string, PCRE compiled incorrect data. This could lead to
4729 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4731 2. Moved to autoconf 2.53 and libtool 1.4.2.
4733 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4734 the special perltest8 script is no longer needed - all the tests can be run
4735 from a single perltest script.
4737 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4738 by \s. It has now been removed in PCRE. This means it isn't recognized as
4739 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4740 class [:space:] *does* include VT, thereby creating a mess.
4742 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4743 space and tab.
4745 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4746 its new features into the main test script, reducing the number of scripts.
4748 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4749 were backward compatible, and made the (?i) apply to the whole pattern, as if
4750 /i were given. Now it behaves more logically, and applies the option setting
4751 only to what follows. PCRE has been changed to follow suit. However, if it
4752 finds options settings right at the start of the pattern, it extracts them into
4753 the global options, as before. Thus, they show up in the info data.
4755 8. Added support for the \Q...\E escape sequence. Characters in between are
4756 treated as literals. This is slightly different from Perl in that $ and @ are
4757 also handled as literals inside the quotes. In Perl, they will cause variable
4758 interpolation. Note the following examples:
4760 Pattern PCRE matches Perl matches
4762 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4763 \Qabc\$xyz\E abc\$xyz abc\$xyz
4764 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4766 For compatibility with Perl, \Q...\E sequences are recognized inside character
4767 classes as well as outside them.
4769 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4770 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4771 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4772 signed/unsigned warnings.
4774 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4775 option for pcretest, so I've replaced it by a simple function that does just
4776 that job.
4778 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4779 "pcregrep -".
4781 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4782 Java package. This provides some syntactic sugar for simple cases of what my
4783 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4784 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4785 item, you can use this simplified notation. Note that only makes sense with
4786 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4787 greediness, whatever the setting of the PCRE_UNGREEDY option.
4789 13. A change of greediness default within a pattern was not taking effect at
4790 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4791 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4792 was abstracted outside.
4794 14. PCRE now supports the \G assertion. It is true when the current matching
4795 position is at the start point of the match. This differs from \A when the
4796 starting offset is non-zero. Used with the /g option of pcretest (or similar
4797 code), it works in the same way as it does for Perl's /g option. If all
4798 alternatives of a regex begin with \G, the expression is anchored to the start
4799 match position, and the "anchored" flag is set in the compiled expression.
4801 15. Some bugs concerning the handling of certain option changes within patterns
4802 have been fixed. These applied to options other than (?ims). For example,
4803 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4804 been the other way round. Some of this was related to change 7 above.
4806 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4807 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4808 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4809 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4811 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4812 mode. Unlike ".", it always matches newline, whatever the setting of
4813 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4814 assertions. Perl allows it, but it doesn't (in general) work because it can't
4815 calculate the length of the lookbehind. At least, that's the case for Perl
4816 5.8.0 - I've been told they are going to document that it doesn't work in
4817 future.
4819 18. Added an error diagnosis for escapes that PCRE does not support: these are
4820 \L, \l, \N, \P, \p, \U, \u, and \X.
4822 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4823 reading past the end of the pattern in cases such as /[abcd/.
4825 20. PCRE was getting more memory than necessary for patterns with classes that
4826 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4828 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4829 compiling PCRE for use with Virtual Pascal.
4831 22. Small fix to the Makefile to make it work properly if the build is done
4832 outside the source tree.
4834 23. Added a new extension: a condition to go with recursion. If a conditional
4835 subpattern starts with (?(R) the "true" branch is used if recursion has
4836 happened, whereas the "false" branch is used only at the top level.
4838 24. When there was a very long string of literal characters (over 255 bytes
4839 without UTF support, over 250 bytes with UTF support), the computation of how
4840 much memory was required could be incorrect, leading to segfaults or other
4841 strange effects.
4843 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4844 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4845 there was a subsequent back reference to those brackets. This meant that, for
4846 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4847 possible to check for precisely this case. All we can do is abandon the
4848 optimization if .* occurs inside capturing brackets when there are any back
4849 references whatsoever. (See below for a better fix that came later.)
4851 26. The handling of the optimization for finding the first character of a
4852 non-anchored pattern, and for finding a character that is required later in the
4853 match were failing in some cases. This didn't break the matching; it just
4854 failed to optimize when it could. The way this is done has been re-implemented.
4856 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4858 28. Added a new feature that provides some of the functionality that Perl
4859 provides with (?{...}). The facility is termed a "callout". The way it is done
4860 in PCRE is for the caller to provide an optional function, by setting
4861 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4862 global variable. By default it is unset, which disables all calling out. To get
4863 the function called, the regex must include (?C) at appropriate points. This
4864 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4865 This provides a means of identifying different callout points. When PCRE
4866 reaches such a point in the regex, if pcre_callout has been set, the external
4867 function is called. It is provided with data in a structure called
4868 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4869 matching continues; if it returns a non-zero value, the match at the current
4870 point fails. However, backtracking will occur if possible. [This was changed
4871 later and other features added - see item 49 below.]
4873 29. pcretest is upgraded to test the callout functionality. It provides a
4874 callout function that displays information. By default, it shows the start of
4875 the match and the current position in the text. There are some new data escapes
4876 to vary what happens:
4878 \C+ in addition, show current contents of captured substrings
4879 \C- do not supply a callout function
4880 \C!n return 1 when callout number n is reached
4881 \C!n!m return 1 when callout number n is reached for the mth time
4883 30. If pcregrep was called with the -l option and just a single file name, it
4884 output "<stdin>" if a match was found, instead of the file name.
4886 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4887 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4888 pcre_exec(). This saves a malloc/free per call. The default value of
4889 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4890 when configuring.
4892 32. The default maximum size of a compiled pattern is 64K. There have been a
4893 few cases of people hitting this limit. The code now uses macros to handle the
4894 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4895 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4896 configuring. Tests 2 and 5 work only with 2-byte links because they output
4897 debugging information about compiled patterns.
4899 33. Internal code re-arrangements:
4901 (a) Moved the debugging function for printing out a compiled regex into
4902 its own source file (printint.c) and used #include to pull it into
4903 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4904 separate copies.
4906 (b) Defined the list of op-code names for debugging as a macro in
4907 internal.h so that it is next to the definition of the opcodes.
4909 (c) Defined a table of op-code lengths for simpler skipping along compiled
4910 code. This is again a macro in internal.h so that it is next to the
4911 definition of the opcodes.
4913 34. Added support for recursive calls to individual subpatterns, along the
4914 lines of Robin Houston's patch (but implemented somewhat differently).
4916 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4917 allow it to read and process whole directories in Win32. This code was
4918 contributed by Lionel Fourquaux; it has not been tested by me.
4920 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4921 used to name a group. Names consist of alphanumerics and underscores, and must
4922 be unique. Back references use the syntax (?P=name) and recursive calls use
4923 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4924 numbers. The function pcre_fullinfo() can be used after compilation to extract
4925 a name/number map. There are three relevant calls:
4927 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4928 PCRE_INFO_NAMECOUNT yields the number of entries
4929 PCRE_INFO_NAMETABLE yields a pointer to the map.
4931 The map is a vector of fixed-size entries. The size of each entry depends on
4932 the length of the longest name used. The first two bytes of each entry are the
4933 group number, most significant byte first. There follows the corresponding
4934 name, zero terminated. The names are in alphabetical order.
4936 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4937 case instead of 255. Making it the same both with and without UTF-8 support
4938 means that the same test output works with both.
4940 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4941 calling malloc() with a zero argument.
4943 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4944 optimization. I've improved things by keeping a bitmap of backreferences with
4945 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4946 fact referenced, the optimization can be applied. It is unlikely that a
4947 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4948 the match to follow \n) will appear inside brackets with a number greater than
4949 31, but if it does, any back reference > 31 suppresses the optimization.
4951 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4952 of disabling numbered capturing parentheses. Any opening parenthesis that is
4953 not followed by ? behaves as if it were followed by ?: but named parentheses
4954 can still be used for capturing (and they will acquire numbers in the usual
4955 way).
4957 41. Redesigned the return codes from the match() function into yes/no/error so
4958 that errors can be passed back from deep inside the nested calls. A malloc
4959 failure while inside a recursive subpattern call now causes the
4960 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4962 42. It is now possible to set a limit on the number of times the match()
4963 function is called in a call to pcre_exec(). This facility makes it possible to
4964 limit the amount of recursion and backtracking, though not in a directly
4965 obvious way, because the match() function is used in a number of different
4966 circumstances. The count starts from zero for each position in the subject
4967 string (for non-anchored patterns). The default limit is, for compatibility, a
4968 large number, namely 10 000 000. You can change this in two ways:
4970 (a) When configuring PCRE before making, you can use --with-match-limit=n
4971 to set a default value for the compiled library.
4973 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4974 a different value is set. See 45 below.
4976 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4978 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4979 of things that can be changed at compile time. The first argument specifies
4980 what is wanted and the second points to where the information is to be placed.
4981 The current list of available information is:
4985 The output is an integer that is set to one if UTF-8 support is available;
4986 otherwise it is set to zero.
4990 The output is an integer that it set to the value of the code that is used for
4991 newline. It is either LF (10) or CR (13).
4995 The output is an integer that contains the number of bytes used for internal
4996 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5000 The output is an integer that contains the threshold above which the POSIX
5001 interface uses malloc() for output vectors. See item 31 above.
5005 The output is an unsigned integer that contains the default limit of the number
5006 of match() calls in a pcre_exec() execution. See 42 above.
5008 44. pcretest has been upgraded by the addition of the -C option. This causes it
5009 to extract all the available output from the new pcre_config() function, and to
5010 output it. The program then exits immediately.
5012 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5013 order to support additional features. One way would have been to define
5014 pcre_exec2() (for example) with extra arguments, but this would not have been
5015 extensible, and would also have required all calls to the original function to
5016 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5017 is used for passing in "extra" data from pcre_study().
5019 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5020 contains the following fields:
5022 flags a bitmap indicating which of the following fields are set
5023 study_data opaque data from pcre_study()
5024 match_limit a way of specifying a limit on match() calls for a specific
5025 call to pcre_exec()
5026 callout_data data for callouts (see 49 below)
5028 The flag bits are also defined in pcre.h, and are
5034 The pcre_study() function now returns one of these new pcre_extra blocks, with
5035 the actual study data pointed to by the study_data field, and the
5036 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5037 before. That is, this change is entirely upwards-compatible and requires no
5038 change to existing code.
5040 If you want to pass in additional data to pcre_exec(), you can either place it
5041 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5042 block.
5044 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5045 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5046 times with different match limits, until it finds the minimum value needed for
5047 pcre_exec() to complete. The value is then output. This can be instructive; for
5048 most simple matches the number is quite small, but for pathological cases it
5049 gets very large very quickly.
5051 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5052 returns the size of the data block pointed to by the study_data field in a
5053 pcre_extra block, that is, the value that was passed as the argument to
5054 pcre_malloc() when PCRE was getting memory in which to place the information
5055 created by pcre_study(). The fourth argument should point to a size_t variable.
5056 pcretest has been extended so that this information is shown after a successful
5057 pcre_study() call when information about the compiled regex is being displayed.
5059 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5060 because what follows is always an absolute path. (Later: it turns out that this
5061 is more than cosmetic for MinGW, because it doesn't like empty path
5062 components.)
5064 49. Some changes have been made to the callout feature (see 28 above):
5066 (i) A callout function now has three choices for what it returns:
5068 0 => success, carry on matching
5069 > 0 => failure at this point, but backtrack if possible
5070 < 0 => serious error, return this value from pcre_exec()
5072 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5073 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5074 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5075 use by callout functions. It will never be used by PCRE itself.
5077 (ii) The pcre_extra structure (see 45 above) has a void * field called
5078 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5079 pcre_callout_block structure has a field of the same name. The contents of
5080 the field passed in the pcre_extra structure are passed to the callout
5081 function in the corresponding field in the callout block. This makes it
5082 easier to use the same callout-containing regex from multiple threads. For
5083 testing, the pcretest program has a new data escape
5085 \C*n pass the number n (may be negative) as callout_data
5087 If the callout function in pcretest receives a non-zero value as
5088 callout_data, it returns that value.
5090 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5091 there were some redundant $(CFLAGS) in commands that are now specified as
5092 $(LINK), which already includes $(CFLAGS).
5094 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5095 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5096 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5097 one-byte characters throughout. Note that case-insensitive matching applies
5098 only to characters whose values are less than 256. PCRE doesn't support the
5099 notion of cases for higher-valued characters.
5101 (i) A character class whose characters are all within 0-255 is handled as
5102 a bit map, and the map is inverted for negative classes. Previously, a
5103 character > 255 always failed to match such a class; however it should
5104 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5106 (ii) A negated character class with a single character < 255 is coded as
5107 "not this character" (OP_NOT). This wasn't working properly when the test
5108 character was multibyte, either singly or repeated.
5110 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5111 mode, for example: \x{100}{2,3}.
5113 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5114 singly or repeated) now correctly test multibyte characters. However,
5115 PCRE doesn't recognize any characters with values greater than 255 as
5116 digits, spaces, or word characters. Such characters always match \D, \S,
5117 and \W, and never match \d, \s, or \w.
5119 (v) Classes may now contain characters and character ranges with values
5120 greater than 255. For example: [ab\x{100}-\x{400}].
5122 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5123 PCRE in UTF-8 mode.
5125 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5126 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5127 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5128 value.)
5130 53. The single man page has become too large. I have therefore split it up into
5131 a number of separate man pages. These also give rise to individual HTML pages;
5132 these are now put in a separate directory, and there is an index.html page that
5133 lists them all. Some hyperlinking between the pages has been installed.
5135 54. Added convenience functions for handling named capturing parentheses.
5137 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5138 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5139 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5140 are faulted.
5142 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5143 calling configure. These values are used when compiling the dftables.c program
5144 which is run to generate the source of the default character tables. They
5145 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5146 you will need to set these values.
5148 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5151 Version 3.9 02-Jan-02
5152 ---------------------
5154 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5156 2. If --disable-static was given, the building process failed when trying to
5157 build pcretest and pcregrep. (For some reason it was using libtool to compile
5158 them, which is not right, as they aren't part of the library.)
5161 Version 3.8 18-Dec-01
5162 ---------------------
5164 1. The experimental UTF-8 code was completely screwed up. It was packing the
5165 bytes in the wrong order. How dumb can you get?
5168 Version 3.7 29-Oct-01
5169 ---------------------
5171 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5172 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5173 this didn't happen under Solaris 8, where I normally test things.
5175 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5176 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5177 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5178 this fix an hour or so after the initial 3.7 release.)
5181 Version 3.6 23-Oct-01
5182 ---------------------
5184 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5185 offsets passed as NULL with zero offset count.
5187 2. The config.guess and config.sub files had not been updated when I moved to
5188 the latest autoconf.
5191 Version 3.5 15-Aug-01
5192 ---------------------
5194 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5195 had been forgotten.
5197 2. By using declared but undefined structures, we can avoid using "void"
5198 definitions in pcre.h while keeping the internal definitions of the structures
5199 private.
5201 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5202 user point of view, this means that both static and shared libraries are built
5203 by default, but this can be individually controlled. More of the work of
5204 handling this static/shared cases is now inside libtool instead of PCRE's make
5205 file.
5207 4. The pcretest utility is now installed along with pcregrep because it is
5208 useful for users (to test regexs) and by doing this, it automatically gets
5209 relinked by libtool. The documentation has been turned into a man page, so
5210 there are now .1, .txt, and .html versions in /doc.
5212 5. Upgrades to pcregrep:
5213 (i) Added long-form option names like gnu grep.
5214 (ii) Added --help to list all options with an explanatory phrase.
5215 (iii) Added -r, --recursive to recurse into sub-directories.
5216 (iv) Added -f, --file to read patterns from a file.
5218 6. pcre_exec() was referring to its "code" argument before testing that
5219 argument for NULL (and giving an error if it was NULL).
5221 7. Upgraded Makefile.in to allow for compiling in a different directory from
5222 the source directory.
5224 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5225 options bits, the pointer it was passed was to an int instead of to an unsigned
5226 long int. This mattered only on 64-bit systems.
5228 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5229 generated) instead of pcre.in, which it its source. Also made the same change
5230 in several of the .c files.
5232 10. A new release of gcc defines printf() as a macro, which broke pcretest
5233 because it had an ifdef in the middle of a string argument for printf(). Fixed
5234 by using separate calls to printf().
5236 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5237 script, to force use of CR or LF instead of \n in the source. On non-Unix
5238 systems, the value can be set in config.h.
5240 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5241 absolute limit. Changed the text of the error message to make this clear, and
5242 likewise updated the man page.
5244 13. The limit of 99 on the number of capturing subpatterns has been removed.
5245 The new limit is 65535, which I hope will not be a "real" limit.
5248 Version 3.4 22-Aug-00
5249 ---------------------
5251 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5253 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5256 Version 3.3 01-Aug-00
5257 ---------------------
5259 1. If an octal character was given, but the value was greater than \377, it
5260 was not getting masked to the least significant bits, as documented. This could
5261 lead to crashes in some systems.
5263 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5264 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5266 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5267 These just pass their arguments on to (pcre_free)(), but they are provided
5268 because some uses of PCRE bind it to non-C systems that can call its functions,
5269 but cannot call free() or pcre_free() directly.
5271 4. Add "make test" as a synonym for "make check". Corrected some comments in
5272 the Makefile.
5274 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5275 Makefile.
5277 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5278 command called pgrep for grepping around the active processes.
5280 7. Added the beginnings of support for UTF-8 character strings.
5282 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5283 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5284 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5285 out for the ar command.)
5288 Version 3.2 12-May-00
5289 ---------------------
5291 This is purely a bug fixing release.
5293 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5294 of ZA. This was just one example of several cases that could provoke this bug,
5295 which was introduced by change 9 of version 2.00. The code for breaking
5296 infinite loops after an iteration that matches an empty string was't working
5297 correctly.
5299 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5300 when matched against abbab (for example). After matching an empty string, it
5301 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5302 caused it to match further down the string than it should.
5304 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5305 was there because it doesn't seem to be needed, and it causes trouble on some
5306 systems, as it is not a Standard C header. It has been removed.
5308 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5309 were reported on the Macintosh. The changes were from
5311 while ((c = *(++ptr)) != 0 && c != '\n');
5312 to
5313 while ((c = *(++ptr)) != 0 && c != '\n') ;
5315 Totally extraordinary, but if that's what it takes...
5317 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5318 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5319 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5320 assumes the way PCRE uses memmove() (always moving upwards).
5322 6. PCRE is being used in one environment where strchr() is not available. There
5323 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5324 faster code anyway.
5327 Version 3.1 09-Feb-00
5328 ---------------------
5330 The only change in this release is the fixing of some bugs in Makefile.in for
5331 the "install" target:
5333 (1) It was failing to install pcreposix.h.
5335 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5338 Version 3.0 01-Feb-00
5339 ---------------------
5341 1. Add support for the /+ modifier to perltest (to output $` like it does in
5342 pcretest).
5344 2. Add support for the /g modifier to perltest.
5346 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5347 matches null strings.
5349 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5350 pattern. Perl treats empty patterns specially - it reuses the most recent
5351 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5352 effect.
5354 5. The POSIX interface was broken in that it was just handing over the POSIX
5355 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5356 required a bigger vector, with some working space on the end. This means that
5357 the POSIX wrapper now has to get and free some memory, and copy the results.
5359 6. Added some simple autoconf support, placing the test data and the
5360 documentation in separate directories, re-organizing some of the
5361 information files, and making it build pcre-config (a GNU standard). Also added
5362 libtool support for building PCRE as a shared library, which is now the
5363 default.
5365 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5366 09 are not valid octal constants. Single digits will be used for minor values
5367 less than 10.
5369 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5370 existing programs that set these in the POSIX interface can use PCRE without
5371 modification.
5373 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5374 return all that pcre_info() returns, plus additional data. The pcre_info()
5375 function is retained for compatibility, but is considered to be obsolete.
5377 10. Added experimental recursion feature (?R) to handle one common case that
5378 Perl 5.6 will be able to do with (?p{...}).
5380 11. Added support for POSIX character classes like [:alpha:], which Perl is
5381 adopting.
5384 Version 2.08 31-Aug-99
5385 ----------------------
5387 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5388 trying to match at the startoffset position, but instead was moving forward to
5389 the next newline as if a previous match had failed.
5391 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5392 and could get into a loop if a null string was matched other than at the start
5393 of the subject.
5395 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5396 be distinguished at compile time, and for completeness also added PCRE_DATE.
5398 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5399 in GnuWin32 environments.
5402 Version 2.07 29-Jul-99
5403 ----------------------
5405 1. The documentation is now supplied in plain text form and HTML as well as in
5406 the form of man page sources.
5408 2. C++ compilers don't like assigning (void *) values to other pointer types.
5409 In particular this affects malloc(). Although there is no problem in Standard
5410 C, I've put in casts to keep C++ compilers happy.
5412 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5413 should be (const char *).
5415 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5416 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5417 However, I haven't made this a standard facility. The documentation doesn't
5418 mention it, and the Makefile doesn't support it.
5420 5. The Makefile now contains an "install" target, with editable destinations at
5421 the top of the file. The pcretest program is not installed.
5423 6. pgrep -V now gives the PCRE version number and date.
5425 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5426 causing the entire string to be ignored, instead of just the last character.
5428 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5429 non-matching string, it can take a very, very long time, even for strings of
5430 quite modest length, because of the nested recursion. PCRE now does better in
5431 some of these cases. It does this by remembering the last required literal
5432 character in the pattern, and pre-searching the subject to ensure it is present
5433 before running the real match. In other words, it applies a heuristic to detect
5434 some types of certain failure quickly, and in the above example, if presented
5435 with a string that has no trailing " it gives "no match" very quickly.
5437 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5438 other alternatives are tried instead.
5441 Version 2.06 09-Jun-99
5442 ----------------------
5444 1. Change pcretest's output for amount of store used to show just the code
5445 space, because the remainder (the data block) varies in size between 32-bit and
5446 64-bit systems.
5448 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5449 start matching at. This allows lookbehinds to work when searching for multiple
5450 occurrences in a string.
5452 3. Added additional options to pcretest for testing multiple occurrences:
5454 /+ outputs the rest of the string that follows a match
5455 /g loops for multiple occurrences, using the new startoffset argument
5456 /G loops for multiple occurrences by passing an incremented pointer
5458 4. PCRE wasn't doing the "first character" optimization for patterns starting
5459 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5460 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5461 the letter 'x'. On long subject strings, this gives a significant speed-up.
5464 Version 2.05 21-Apr-99
5465 ----------------------
5467 1. Changed the type of magic_number from int to long int so that it works
5468 properly on 16-bit systems.
5470 2. Fixed a bug which caused patterns starting with .* not to work correctly
5471 when the subject string contained newline characters. PCRE was assuming
5472 anchoring for such patterns in all cases, which is not correct because .* will
5473 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5474 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5475 must be retried after every newline in the subject.
5478 Version 2.04 18-Feb-99
5479 ----------------------
5481 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5482 computation of the store needed to hold the pattern was incorrect (too large).
5483 If such patterns were nested a few deep, this could multiply and become a real
5484 problem.
5486 2. Added /M option to pcretest to show the memory requirement of a specific
5487 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5489 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5490 compiled in such a way that the backtracking after subsequent failure was
5491 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5492 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5495 Version 2.03 02-Feb-99
5496 ----------------------
5498 1. Fixed typo and small mistake in man page.
5500 2. Added 4th condition (GPL supersedes if conflict) and created separate
5501 LICENCE file containing the conditions.
5503 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5504 Perl, that is the internal \ allows the delimiter to be included in the
5505 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5506 the final delimiter, add \ to the end of the pattern (to test the error).
5508 4. Added the convenience functions for extracting substrings after a successful
5509 match. Updated pcretest to make it able to test these functions.
5512 Version 2.02 14-Jan-99
5513 ----------------------
5515 1. Initialized the working variables associated with each extraction so that
5516 their saving and restoring doesn't refer to uninitialized store.
5518 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5519 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5520 fix the problem.
5522 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5523 calls, and wasn't printing the correct value for compiling calls. Increased the
5524 default value of LOOPREPEAT, and the number of significant figures in the
5525 times.
5527 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5529 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5530 a building problem on Windows NT with a FAT file system.
5533 Version 2.01 21-Oct-98
5534 ----------------------
5536 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5537 to character tables built by pcre_maketables() in the current locale. If NULL
5538 is passed, the default tables are used.
5541 Version 2.00 24-Sep-98
5542 ----------------------
5544 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5545 it any more.
5547 2. Allow quantification of (?>) groups, and make it work correctly.
5549 3. The first character computation wasn't working for (?>) groups.
5551 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5552 end of the subject) and add 5.005's \z, which really does match only at the
5553 very end of the subject.
5555 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5557 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5558 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5559 localized options. All options to pcre_study() were also removed.
5561 7. Add other new features from 5.005:
5563 $(?<= positive lookbehind
5564 $(?<! negative lookbehind
5565 (?imsx-imsx) added the unsetting capability
5566 such a setting is global if at outer level; local otherwise
5567 (?imsx-imsx:) non-capturing groups with option setting
5568 (?(cond)re|re) conditional pattern matching
5570 A backreference to itself in a repeated group matches the previous
5571 captured string.
5573 8. General tidying up of studying (both automatic and via "study")
5574 consequential on the addition of new assertions.
5576 9. As in 5.005, unlimited repeated groups that could match an empty substring
5577 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5578 runtime if any iteration does actually match an empty substring.
5580 10. Include the RunTest script in the distribution.
5582 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5583 discrepancies, some of which were old and were also with respect to 5.004. They
5584 have now been fixed.
5587 Version 1.09 28-Apr-98
5588 ----------------------
5590 1. A negated single character class followed by a quantifier with a minimum
5591 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5592 program crashes, or just wrong answers. This did not apply to negated classes
5593 containing more than one character, or to minima other than one.
5596 Version 1.08 27-Mar-98
5597 ----------------------
5599 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5601 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5602 latter must appear before anything that relies on it in the pattern.
5605 Version 1.07 16-Feb-98
5606 ----------------------
5608 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5609 repeat of a potentially empty string).
5612 Version 1.06 23-Jan-98
5613 ----------------------
5615 1. Added Markus Oberhumer's little patches for C++.
5617 2. Literal strings longer than 255 characters were broken.
5620 Version 1.05 23-Dec-97
5621 ----------------------
5623 1. Negated character classes containing more than one character were failing if
5624 PCRE_CASELESS was set at run time.
5627 Version 1.04 19-Dec-97
5628 ----------------------
5630 1. Corrected the man page, where some "const" qualifiers had been omitted.
5632 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5633 input syntax.
5635 3. Fixed memory leak which occurred when a regex with back references was
5636 matched with an offsets vector that wasn't big enough. The temporary memory
5637 that is used in this case wasn't being freed if the match failed.
5639 4. Tidied pcretest to ensure it frees memory that it gets.
5641 5. Temporary memory was being obtained in the case where the passed offsets
5642 vector was exactly big enough.
5644 6. Corrected definition of offsetof() from change 5 below.
5646 7. I had screwed up change 6 below and broken the rules for the use of
5647 setjmp(). Now fixed.
5650 Version 1.03 18-Dec-97
5651 ----------------------
5653 1. A erroneous regex with a missing opening parenthesis was correctly
5654 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5655 on some systems.
5657 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5658 it was reported that one broken compiler failed on the former because "code" is
5659 also an independent variable.
5661 3. The erroneous regex a[]b caused an array overrun reference.
5663 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5664 fail on data ending with that character. (It was going on too far, and checking
5665 the next character, typically a binary zero.) This was specific to the
5666 optimized code for single-character negative classes.
5668 5. Added a contributed patch from the TIN world which does the following:
5670 + Add an undef for memmove, in case the the system defines a macro for it.
5672 + Add a definition of offsetof(), in case there isn't one. (I don't know
5673 the reason behind this - offsetof() is part of the ANSI standard - but
5674 it does no harm).
5676 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5677 most of the places where whitespace preceded '#'. I have given up and
5678 allowed the remaining 2 cases to be at the margin.
5680 + Rename some variables in pcre to eliminate shadowing. This seems very
5681 pedantic, but does no harm, of course.
5683 6. Moved the call to setjmp() into its own function, to get rid of warnings
5684 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5686 7. Constructs such as \d{8,} were compiling into the equivalent of
5687 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5688 outcome, but in this particular case used more store than had been allocated,
5689 which caused the bug to be discovered because it threw up an internal error.
5691 8. The debugging code in both pcre and pcretest for outputting the compiled
5692 form of a regex was going wrong in the case of back references followed by
5693 curly-bracketed repeats.
5696 Version 1.02 12-Dec-97
5697 ----------------------
5699 1. Typos in pcre.3 and comments in the source fixed.
5701 2. Applied a contributed patch to get rid of places where it used to remove
5702 'const' from variables, and fixed some signed/unsigned and uninitialized
5703 variable warnings.
5705 3. Added the "runtest" target to Makefile.
5707 4. Set default compiler flag to -O2 rather than just -O.
5710 Version 1.01 19-Nov-97
5711 ----------------------
5713 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5714 like /([ab]*)*/, that is, for classes with more than one character in them.
5716 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5717 as /((?>a*))*/ (a PCRE_EXTRA facility).
5720 Version 1.00 18-Nov-97
5721 ----------------------
5723 1. Added compile-time macros to support systems such as SunOS4 which don't have
5724 memmove() or strerror() but have other things that can be used instead.
5726 2. Arranged that "make clean" removes the executables.
5729 Version 0.99 27-Oct-97
5730 ----------------------
5732 1. Fixed bug in code for optimizing classes with only one character. It was
5733 initializing a 32-byte map regardless, which could cause it to run off the end
5734 of the memory it had got.
5736 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5739 Version 0.98 22-Oct-97
5740 ----------------------
5742 1. Fixed bug in code for handling temporary memory usage when there are more
5743 back references than supplied space in the ovector. This could cause segfaults.
5746 Version 0.97 21-Oct-97
5747 ----------------------
5749 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5751 2. Optimized negated single characters not to use a bit map.
5753 3. Brought error texts together as macro definitions; clarified some of them;
5754 fixed one that was wrong - it said "range out of order" when it meant "invalid
5755 escape sequence".
5757 4. Changed some char * arguments to const char *.
5759 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5761 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5762 pcretest.
5765 Version 0.96 16-Oct-97
5766 ----------------------
5768 1. Added a simple "pgrep" utility to the distribution.
5770 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5771 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5772 where "ddd" means "one or more decimal digits".
5774 3. Fixed serious bug. If a pattern had a back reference, but the call to
5775 pcre_exec() didn't supply a large enough ovector to record the related
5776 identifying subpattern, the match always failed. PCRE now remembers the number
5777 of the largest back reference, and gets some temporary memory in which to save
5778 the offsets during matching if necessary, in order to ensure that
5779 backreferences always work.
5781 4. Increased the compatibility with Perl in a number of ways:
5783 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5784 to request this handling. The option can be set at compile or exec time.
5786 (b) $ matches before a terminating newline by default; an option
5787 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5788 mode). The option can be set at compile or exec time.
5790 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5791 the same as Perl's. If the decimal number it represents is less than 10
5792 or there aren't that many previous left capturing parentheses, an octal
5793 escape is read. Inside a character class, it's always an octal escape,
5794 even if it is a single digit.
5796 (d) An escaped but undefined alphabetic character is taken as a literal,
5797 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5798 escapes.
5800 (e) {0} is now permitted. (The previous item is removed from the compiled
5801 pattern).
5803 5. Changed all the names of code files so that the basic parts are no longer
5804 than 10 characters, and abolished the teeny "globals.c" file.
5806 6. Changed the handling of character classes; they are now done with a 32-byte
5807 bit map always.
5809 7. Added the -d and /D options to pcretest to make it possible to look at the
5810 internals of compilation without having to recompile pcre.
5813 Version 0.95 23-Sep-97
5814 ----------------------
5816 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5817 \x20 at the start of a run of normal characters. These were being treated as
5818 real characters, instead of the source characters being re-checked.
5821 Version 0.94 18-Sep-97
5822 ----------------------
5824 1. The functions are now thread-safe, with the caveat that the global variables
5825 containing pointers to malloc() and free() or alternative functions are the
5826 same for all threads.
5828 2. Get pcre_study() to generate a bitmap of initial characters for non-
5829 anchored patterns when this is possible, and use it if passed to pcre_exec().
5832 Version 0.93 15-Sep-97
5833 ----------------------
5835 1. /(b)|(:+)/ was computing an incorrect first character.
5837 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5838 but not actually doing anything yet.
5840 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5841 as Perl does (e.g. [-az] or [az-]).
5843 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5844 all possible positions.
5846 5. Split up into different modules to avoid including unneeded functions in a
5847 compiled binary. However, compile and exec are still in one module. The "study"
5848 function is split off.
5850 6. The character tables are now in a separate module whose source is generated
5851 by an auxiliary program - but can then be edited by hand if required. There are
5852 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5853 toupper() in the code.
5855 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5856 make them global. Abolish the function for setting them, as the caller can now
5857 set them directly.
5860 Version 0.92 11-Sep-97
5861 ----------------------
5863 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5864 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5866 2. Caseless matching was not working in character classes if the characters in
5867 the pattern were in upper case.
5869 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5871 4. Make PCRE_ANCHORED public and accept as a compile option.
5873 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5874 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5875 pass them.
5877 6. Give an error if bad option bits passed at compile or run time.
5879 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5880 pcretest to cause it to pass that flag.
5882 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5883 options, and the first character, if set.
5885 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5888 Version 0.91 10-Sep-97
5889 ----------------------
5891 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5892 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5894 2. PCRE was looping on encountering an indefinitely repeated back reference to
5895 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5896 Perl does - treats the match as successful.
5898 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5