ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1587 - (show annotations)
Sat Aug 8 06:02:16 2015 UTC (4 years, 4 months ago) by zherczeg
File size: 282576 byte(s)
Error occurred while calculating annotation data.
The JIT compiler did not restore the control verb head in case of *THEN control verbs.
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 fuzzer.
117 Version 8.37 28-April-2015
118 --------------------------
120 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
121 for those parentheses to be closed with whatever has been captured so far.
122 However, it was failing to mark any other groups between the hightest
123 capture so far and the currrent group as "unset". Thus, the ovector for
124 those groups contained whatever was previously there. An example is the
125 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
127 2. If an assertion condition was quantified with a minimum of zero (an odd
128 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
130 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
131 unrecognized modifier, a crash could occur.
133 4. An attempt to do global matching in pcretest with a zero-length ovector
134 caused a crash.
136 5. Fixed a memory leak during matching that could occur for a subpattern
137 subroutine call (recursive or otherwise) if the number of captured groups
138 that had to be saved was greater than ten.
140 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
141 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
142 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
144 7. A UTF pattern containing a "not" match of a non-ASCII character and a
145 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
147 8. When a pattern is compiled, it remembers the highest back reference so that
148 when matching, if the ovector is too small, extra memory can be obtained to
149 use instead. A conditional subpattern whose condition is a check on a
150 capture having happened, such as, for example in the pattern
151 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
152 setting the highest backreference number. This mattered only if pcre_exec()
153 was called with an ovector that was too small to hold the capture, and there
154 was no other kind of back reference (a situation which is probably quite
155 rare). The effect of the bug was that the condition was always treated as
156 FALSE when the capture could not be consulted, leading to a incorrect
157 behaviour by pcre_exec(). This bug has been fixed.
159 9. A reference to a duplicated named group (either a back reference or a test
160 for being set in a conditional) that occurred in a part of the pattern where
161 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
162 to be incorrectly calculated, leading to overwriting.
164 10. A mutually recursive set of back references such as (\2)(\1) caused a
165 segfault at study time (while trying to find the minimum matching length).
166 The infinite loop is now broken (with the minimum length unset, that is,
167 zero).
169 11. If an assertion that was used as a condition was quantified with a minimum
170 of zero, matching went wrong. In particular, if the whole group had
171 unlimited repetition and could match an empty string, a segfault was
172 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
173 assertions to be quantified, but not if they are being used as conditions,
174 so the above pattern is faulted by Perl. PCRE has now been changed so that
175 it also rejects such patterns.
177 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
178 failed to allow the zero-repeat case if pcre2_exec() was called with an
179 ovector too small to capture the group.
181 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
182 Red Hat Product Security:
184 (a) A crash if /K and /F were both set with the option to save the compiled
185 pattern.
187 (b) Another crash if the option to print captured substrings in a callout
188 was combined with setting a null ovector, for example \O\C+ as a subject
189 string.
191 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
192 forward reference repeated a large (but limited) number of times within a
193 repeated outer group that has a zero minimum quantifier, caused incorrect
194 code to be compiled, leading to the error "internal error:
195 previously-checked referenced subpattern not found" when an incorrect
196 memory address was read. This bug was reported as "heap overflow",
197 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
198 CVE-2015-2325.
200 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
201 call within a group that also contained a recursive back reference caused
202 incorrect code to be compiled. This bug was reported as "heap overflow",
203 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
204 number CVE-2015-2326.
206 24. Computing the size of the JIT read-only data in advance has been a source
207 of various issues, and new ones are still appear unfortunately. To fix
208 existing and future issues, size computation is eliminated from the code,
209 and replaced by on-demand memory allocation.
211 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
212 adjacent to the end of the range, and the range contained characters with
213 more than one other case, caused incorrect behaviour when compiled in UTF
214 mode. In that example, the range a-j was left out of the class.
216 26. Fix JIT compilation of conditional blocks, which assertion
217 is converted to (*FAIL). E.g: /(?(?!))/.
219 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
220 discovered by the LLVM fuzzer.
222 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
223 when this assertion was used as a condition, for example (?(?!)a|b). In
224 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
225 error about an unsupported item.
227 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
228 possessification code could take exponential time to complete. A recursion
229 depth limit of 1000 has been imposed to limit the resources used by this
230 optimization.
232 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
233 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
234 because \S ensures they are all in the class. The code for doing this was
235 interacting badly with the code for computing the amount of space needed to
236 compile the pattern, leading to a buffer overflow. This bug was discovered
237 by the LLVM fuzzer.
239 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
240 other kinds of group caused stack overflow at compile time. This bug was
241 discovered by the LLVM fuzzer.
243 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
244 between a subroutine call and its quantifier was incorrectly compiled,
245 leading to buffer overflow or other errors. This bug was discovered by the
246 LLVM fuzzer.
248 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
249 assertion after (?(. The code was failing to check the character after
250 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
251 was discovered by the LLVM fuzzer.
253 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
254 a fixed maximum following a group that contains a subroutine reference was
255 incorrectly compiled and could trigger buffer overflow. This bug was
256 discovered by the LLVM fuzzer.
258 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
259 caused a stack overflow instead of the diagnosis of a non-fixed length
260 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
262 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
263 (e.g. /(?<=\Ka)/) could make pcregrep loop.
265 37. There was a similar problem to 36 in pcretest for global matches.
267 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
268 and a subsequent item in the pattern caused a non-match, backtracking over
269 the repeated \X did not stop, but carried on past the start of the subject,
270 causing reference to random memory and/or a segfault. There were also some
271 other cases where backtracking after \C could crash. This set of bugs was
272 discovered by the LLVM fuzzer.
274 39. The function for finding the minimum length of a matching string could take
275 a very long time if mutual recursion was present many times in a pattern,
276 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
277 method has been implemented. This infelicity was discovered by the LLVM
278 fuzzer.
280 40. Static linking against the PCRE library using the pkg-config module was
281 failing on missing pthread symbols.
284 Version 8.36 26-September-2014
285 ------------------------------
287 1. Got rid of some compiler warnings in the C++ modules that were shown up by
288 -Wmissing-field-initializers and -Wunused-parameter.
290 2. The tests for quantifiers being too big (greater than 65535) were being
291 applied after reading the number, and stupidly assuming that integer
292 overflow would give a negative number. The tests are now applied as the
293 numbers are read.
295 3. Tidy code in pcre_exec.c where two branches that used to be different are
296 now the same.
298 4. The JIT compiler did not generate match limit checks for certain
299 bracketed expressions with quantifiers. This may lead to exponential
300 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
301 issue should be resolved now.
303 5. Fixed an issue, which occures when nested alternatives are optimized
304 with table jumps.
306 6. Inserted two casts and changed some ints to size_t in the light of some
307 reported 64-bit compiler warnings (Bugzilla 1477).
309 7. Fixed a bug concerned with zero-minimum possessive groups that could match
310 an empty string, which sometimes were behaving incorrectly in the
311 interpreter (though correctly in the JIT matcher). This pcretest input is
312 an example:
314 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
317 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
318 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
319 for an empty string was breaking the inner loop and carrying on at a lower
320 level, when possessive repeated groups should always return to a higher
321 level as they have no backtrack points in them. The empty string test now
322 occurs at the outer level.
324 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
325 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
327 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
328 doesn't).
330 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
331 to hold a pattern to be incorrectly computed (too small) when there were
332 named back references to duplicated names. This could cause "internal
333 error: code overflow" or "double free or corruption" or other memory
334 handling errors.
336 11. When named subpatterns had the same prefixes, back references could be
337 confused. For example, in this pattern:
339 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
341 the reference to 'Name' was incorrectly treated as a reference to a
342 duplicate name.
344 12. A pattern such as /^s?c/mi8 where the optional character has more than
345 one "other case" was incorrectly compiled such that it would only try to
346 match starting at "c".
348 13. When a pattern starting with \s was studied, VT was not included in the
349 list of possible starting characters; this should have been part of the
350 8.34/18 patch.
352 14. If a character class started [\Qx]... where x is any character, the class
353 was incorrectly terminated at the ].
355 15. If a pattern that started with a caseless match for a character with more
356 than one "other case" was studied, PCRE did not set up the starting code
357 unit bit map for the list of possible characters. Now it does. This is an
358 optimization improvement, not a bug fix.
360 16. The Unicode data tables have been updated to Unicode 7.0.0.
362 17. Fixed a number of memory leaks in pcregrep.
364 18. Avoid a compiler warning (from some compilers) for a function call with
365 a cast that removes "const" from an lvalue by using an intermediate
366 variable (to which the compiler does not object).
368 19. Incorrect code was compiled if a group that contained an internal recursive
369 back reference was optional (had quantifier with a minimum of zero). This
370 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
371 caused segmentation faults because of stack overflows at compile time.
373 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
374 group that is quantified with an indefinite repeat, caused a compile-time
375 loop which used up all the system stack and provoked a segmentation fault.
376 This was not the same bug as 19 above.
378 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
379 Patch by Mike Frysinger.
382 Version 8.35 04-April-2014
383 --------------------------
385 1. A new flag is set, when property checks are present in an XCLASS.
386 When this flag is not set, PCRE can perform certain optimizations
387 such as studying these XCLASS-es.
389 2. The auto-possessification of character sets were improved: a normal
390 and an extended character set can be compared now. Furthermore
391 the JIT compiler optimizes more character set checks.
393 3. Got rid of some compiler warnings for potentially uninitialized variables
394 that show up only when compiled with -O2.
396 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
397 of a match later then the end of the match. The pcretest program was not
398 handling the case sensibly - it was outputting from the start to the next
399 binary zero. It now reports this situation in a message, and outputs the
400 text from the end to the start.
402 5. Fast forward search is improved in JIT. Instead of the first three
403 characters, any three characters with fixed position can be searched.
404 Search order: first, last, middle.
406 6. Improve character range checks in JIT. Characters are read by an inprecise
407 function now, which returns with an unknown value if the character code is
408 above a certain threshold (e.g: 256). The only limitation is that the value
409 must be bigger than the threshold as well. This function is useful when
410 the characters above the threshold are handled in the same way.
412 7. The macros whose names start with RAWUCHAR are placeholders for a future
413 mode in which only the bottom 21 bits of 32-bit data items are used. To
414 make this more memorable for those maintaining the code, the names have
415 been changed to start with UCHAR21, and an extensive comment has been added
416 to their definition.
418 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
419 to the export list in Makefile.am (they were accidentally omitted from the
420 8.34 tarball).
422 9. The informational output from pcretest used the phrase "starting byte set"
423 which is inappropriate for the 16-bit and 32-bit libraries. As the output
424 for "first char" and "need char" really means "non-UTF-char", I've changed
425 "byte" to "char", and slightly reworded the output. The documentation about
426 these values has also been (I hope) clarified.
428 10. Another JIT related optimization: use table jumps for selecting the correct
429 backtracking path, when more than four alternatives are present inside a
430 bracket.
432 11. Empty match is not possible, when the minimum length is greater than zero,
433 and there is no \K in the pattern. JIT should avoid empty match checks in
434 such cases.
436 12. In a caseless character class with UCP support, when a character with more
437 than one alternative case was not the first character of a range, not all
438 the alternative cases were added to the class. For example, s and \x{17f}
439 are both alternative cases for S: the class [RST] was handled correctly,
440 but [R-T] was not.
442 13. The configure.ac file always checked for pthread support when JIT was
443 enabled. This is not used in Windows, so I have put this test inside a
444 check for the presence of windows.h (which was already tested for).
446 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
447 The algorithm provides a way to skip certain starting offsets, and usually
448 faster than linear prefix searches.
450 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
451 as for 'fr_FR' and 'french'. For some reason, however, it then used the
452 Windows-specific input and output files, which have 'french' screwed in.
453 So this could never have worked. One of the problems with locales is that
454 they aren't always the same. I have now updated RunTest so that it checks
455 the output of the locale test (test 3) against three different output
456 files, and it allows the test to pass if any one of them matches. With luck
457 this should make the test pass on some versions of Solaris where it was
458 failing. Because of the uncertainty, the script did not used to stop if
459 test 3 failed; it now does. If further versions of a French locale ever
460 come to light, they can now easily be added.
462 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
463 there was a message during ./configure, but it did not stop. This now
464 provokes an error. The invalid example in README has been corrected.
465 If a value less than the minimum is given, the minimum value has always
466 been used, but now a warning is given.
468 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
469 was a bug in the test system, which is now fixed. Also, the list of various
470 configurations that are tested for each release did not have one with both
471 16/32 bits and --enable-bar-anycrlf. It now does.
473 18. pcretest was missing "-C bsr" for displaying the \R default setting.
475 19. Little endian PowerPC systems are supported now by the JIT compiler.
477 20. The fast forward newline mechanism could enter to an infinite loop on
478 certain invalid UTF-8 input. Although we don't support these cases
479 this issue can be fixed by a performance optimization.
481 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
482 not take account if existing stack usage. There is now a new global
483 variable called pcre_stack_guard that can be set to point to an external
484 function to check stack availability. It is called at the start of
485 processing every parenthesized group.
487 22. A typo in the code meant that in ungreedy mode the max/min qualifier
488 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
489 not match "ab".
491 23. When UTF was disabled, the JIT program reported some incorrect compile
492 errors. These messages are silenced now.
494 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
495 compiler.
497 25. Change all the temporary files used in RunGrepTest to be different to those
498 used by RunTest so that the tests can be run simultaneously, for example by
499 "make -j check".
502 Version 8.34 15-December-2013
503 -----------------------------
505 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
506 executable memory. Patch inspired by Carsten Klein.
508 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
509 this macro is never tested and has no effect, because the work to support
510 coverage involves only compiling and linking options and special targets in
511 the Makefile. The comment in config.h implied that defining the macro would
512 enable coverage support, which is totally false. There was also support for
513 setting this macro in the CMake files (my fault, I just copied it from
514 configure). SUPPORT_GCOV has now been removed.
516 3. Make a small performance improvement in strlen16() and strlen32() in
517 pcretest.
519 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
520 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
521 these cases). There was also one in pcretest.c.
523 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
525 6. In UTF mode, the code for checking whether a group could match an empty
526 string (which is used for indefinitely repeated groups to allow for
527 breaking an infinite loop) was broken when the group contained a repeated
528 negated single-character class with a character that occupied more than one
529 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
530 UTF-8 mode). The effect was undefined: the group might or might not be
531 deemed as matching an empty string, or the program might have crashed.
533 7. The code for checking whether a group could match an empty string was not
534 recognizing that \h, \H, \v, \V, and \R must match a character.
536 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
537 an empty string. If it can, pcretest shows this in its information output.
539 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
540 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
541 matched by pcre_exec() without using JIT:
543 (a) If the rest of the pattern did not match after a maximal run of
544 grapheme clusters, the code for backing up to try with fewer of them
545 did not always back up over a full grapheme when characters that do not
546 have the modifier quality were involved, e.g. Hangul syllables.
548 (b) If the match point in a subject started with modifier character, and
549 there was no match, the code could incorrectly back up beyond the match
550 point, and potentially beyond the first character in the subject,
551 leading to a segfault or an incorrect match result.
553 10. A conditional group with an assertion condition could lead to PCRE
554 recording an incorrect first data item for a match if no other first data
555 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
556 first data item, and therefore matched "ca" after "c" instead of at the
557 start.
559 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
560 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
562 12. The source of pcregrep now includes z/OS-specific code so that it can be
563 compiled for z/OS as part of the special z/OS distribution.
565 13. Added the -T and -TM options to pcretest.
567 14. The code in pcre_compile.c for creating the table of named capturing groups
568 has been refactored. Instead of creating the table dynamically during the
569 actual compiling pass, the information is remembered during the pre-compile
570 pass (on the stack unless there are more than 20 named groups, in which
571 case malloc() is used) and the whole table is created before the actual
572 compile happens. This has simplified the code (it is now nearly 150 lines
573 shorter) and prepared the way for better handling of references to groups
574 with duplicate names.
576 15. A back reference to a named subpattern when there is more than one of the
577 same name now checks them in the order in which they appear in the pattern.
578 The first one that is set is used for the reference. Previously only the
579 first one was inspected. This change makes PCRE more compatible with Perl.
581 16. Unicode character properties were updated from Unicode 6.3.0.
583 17. The compile-time code for auto-possessification has been refactored, based
584 on a patch by Zoltan Herczeg. It now happens after instead of during
585 compilation. The code is cleaner, and more cases are handled. The option
586 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
587 options in pcretest are provided to set it. It can also be set by
588 (*NO_AUTO_POSSESS) at the start of a pattern.
590 18. The character VT has been added to the default ("C" locale) set of
591 characters that match \s and are generally treated as white space,
592 following this same change in Perl 5.18. There is now no difference between
593 "Perl space" and "POSIX space". Whether VT is treated as white space in
594 other locales depends on the locale.
596 19. The code for checking named groups as conditions, either for being set or
597 for being recursed, has been refactored (this is related to 14 and 15
598 above). Processing unduplicated named groups should now be as fast at
599 numerical groups, and processing duplicated groups should be faster than
600 before.
602 20. Two patches to the CMake build system, by Alexander Barkov:
604 (1) Replace the "source" command by "." in CMakeLists.txt because
605 "source" is a bash-ism.
607 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
608 without these the CMake build does not work on Solaris.
610 21. Perl has changed its handling of \8 and \9. If there is no previously
611 encountered capturing group of those numbers, they are treated as the
612 literal characters 8 and 9 instead of a binary zero followed by the
613 literals. PCRE now does the same.
615 22. Following Perl, added \o{} to specify codepoints in octal, making it
616 possible to specify values greater than 0777 and also making them
617 unambiguous.
619 23. Perl now gives an error for missing closing braces after \x{... instead of
620 treating the string as literal. PCRE now does the same.
622 24. RunTest used to grumble if an inappropriate test was selected explicitly,
623 but just skip it when running all tests. This make it awkward to run ranges
624 of tests when one of them was inappropriate. Now it just skips any
625 inappropriate tests, as it always did when running all tests.
627 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
628 character types such as \d or \w, too many callouts were inserted, and the
629 data that they returned was rubbish.
631 26. In UCP mode, \s was not matching two of the characters that Perl matches,
632 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
633 were matched by \h. The code has now been refactored so that the lists of
634 the horizontal and vertical whitespace characters used for \h and \v (which
635 are defined only in one place) are now also used for \s.
637 27. Add JIT support for the 64 bit TileGX architecture.
638 Patch by Jiong Wang (Tilera Corporation).
640 28. Possessive quantifiers for classes (both explicit and automatically
641 generated) now use special opcodes instead of wrapping in ONCE brackets.
643 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
644 (because it's meaningless), this was not happening when PCRE_CASELESS was
645 set. Not wrong, but inefficient.
647 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
648 properties for \w, \d, etc) is present in a test regex. Otherwise if the
649 test contains no characters greater than 255, Perl doesn't realise it
650 should be using Unicode semantics.
652 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
653 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
654 does in Unicode mode.
656 32. Added the "forbid" facility to pcretest so that putting tests into the
657 wrong test files can sometimes be quickly detected.
659 33. There is now a limit (default 250) on the depth of nesting of parentheses.
660 This limit is imposed to control the amount of system stack used at compile
661 time. It can be changed at build time by --with-parens-nest-limit=xxx or
662 the equivalent in CMake.
664 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
665 errors. Perl warns for these when in warning mode, but PCRE has no facility
666 for giving warnings.
668 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
669 However, this was not working for (?!) because it is optimized to (*FAIL),
670 for which PCRE does not allow quantifiers. The optimization is now disabled
671 when a quantifier follows (?!). I can't see any use for this, but it makes
672 things uniform.
674 36. Perl no longer allows group names to start with digits, so I have made this
675 change also in PCRE. It simplifies the code a bit.
677 37. In extended mode, Perl ignores spaces before a + that indicates a
678 possessive quantifier. PCRE allowed a space before the quantifier, but not
679 before the possessive +. It now does.
681 38. The use of \K (reset reported match start) within a repeated possessive
682 group such as (a\Kb)*+ was not working.
684 40. Document that the same character tables must be used at compile time and
685 run time, and that the facility to pass tables to pcre_exec() and
686 pcre_dfa_exec() is for use only with saved/restored patterns.
688 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
689 features for Builds with MSVC:
691 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
692 builds). The .rc files can be used to set FileDescription and many other
693 attributes.
695 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
696 This allows higher-level build scripts which want .pdb files to avoid
697 hard-coding the exact files needed."
699 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
700 mean "start of word" and "end of word", respectively, as a transition aid.
702 43. A minimizing repeat of a class containing codepoints greater than 255 in
703 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
704 compiled to use the heap for recursion.
706 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
707 is configured.
710 Version 8.33 28-May-2013
711 ------------------------
713 1. Added 'U' to some constants that are compared to unsigned integers, to
714 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
715 variables that are added to signed variables, to ensure the result is
716 signed and can be negated.
718 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
719 CMake config files.
721 3. Revise the creation of config.h.generic so that all boolean macros are
722 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
723 overriding via -D on the command line possible.
725 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
726 to unsigned int is reported to make a quite noticeable speed difference in
727 a specific Windows environment. Testing on Linux did also appear to show
728 some benefit (and it is clearly not harmful). Also fixed the definition of
729 Xop which should be unsigned.
731 5. Related to (4), changing the definition of the intermediate variable cc
732 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
733 improvements.
735 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
736 unnecessary spaces.
738 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
739 and later.
741 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
743 9. Optimizing fast_forward_start_bits in JIT.
745 10. Adding support for callouts in JIT, and fixing some issues revealed
746 during this work. Namely:
748 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
750 (b) Minimum length was not checked before the matching is started.
752 11. The value of capture_last that is passed to callouts was incorrect in some
753 cases when there was a capture on one path that was subsequently abandoned
754 after a backtrack. Also, the capture_last value is now reset after a
755 recursion, since all captures are also reset in this case.
757 12. The interpreter no longer returns the "too many substrings" error in the
758 case when an overflowing capture is in a branch that is subsequently
759 abandoned after a backtrack.
761 13. In the pathological case when an offset vector of size 2 is used, pcretest
762 now prints out the matched string after a yield of 0 or 1.
764 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
765 Only supported by the JIT compiler at the moment.
767 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
769 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
770 offset of the starting point of the matching process, provided the offsets
771 vector is large enough.
773 17. The \A escape now records a lookbehind value of 1, though its execution
774 does not actually inspect the previous character. This is to ensure that,
775 in partial multi-segment matching, at least one character from the old
776 segment is retained when a new segment is processed. Otherwise, if there
777 are no lookbehinds in the pattern, \A might match incorrectly at the start
778 of a new segment.
780 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
782 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
783 gives some modest performance improvement in 8-bit mode.
785 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
786 be expressed in certain programming languages using Universal Character
787 Names.
789 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
790 which points out that "non characters" are not "characters that may not
791 appear in Unicode strings" but rather "characters that are reserved for
792 internal use and have only local meaning".
794 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
795 there was a conditional group that depended on an assertion, if the
796 assertion was false, the callout that immediately followed the alternation
797 in the condition was skipped when pcre_exec() was used for matching.
799 23. Allow an explicit callout to be inserted before an assertion that is the
800 condition for a conditional group, for compatibility with automatic
801 callouts, which always insert a callout at this point.
803 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
804 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
806 25. (*PRUNE) is now supported by the JIT compiler.
808 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
810 27. Fix the case where there are two or more SKIPs with arguments that may be
811 ignored.
813 28. (*SKIP) is now supported by the JIT compiler.
815 29. (*THEN) is now supported by the JIT compiler.
817 30. Update RunTest with additional test selector options.
819 31. The way PCRE handles backtracking verbs has been changed in two ways.
821 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
822 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
823 backtracking. In some cases this makes it more Perl-compatible, but Perl's
824 rather obscure rules do not always do the same thing.
826 (2) Previously, backtracking verbs were confined within assertions. This is
827 no longer the case for positive assertions, except for (*ACCEPT). Again,
828 this sometimes improves Perl compatibility, and sometimes does not.
830 32. A number of tests that were in test 2 because Perl did things differently
831 have been moved to test 1, because either Perl or PCRE has changed, and
832 these tests are now compatible.
834 32. Backtracking control verbs are now handled in the same way in JIT and
835 interpreter.
837 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
838 contained a forward subroutine reference caused a compile error.
840 34. Auto-detect and optimize limited repetitions in JIT.
842 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
843 blocking (*UTF) etc.
845 36. In the interpreter, maximizing pattern repetitions for characters and
846 character types now use tail recursion, which reduces stack usage.
848 37. The value of the max lookbehind was not correctly preserved if a compiled
849 and saved regex was reloaded on a host of different endianness.
851 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
852 of the compiled pattern block, expand the flags field from 16 to 32 bits
853 because it was almost full.
855 39. Try madvise first before posix_madvise.
857 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
858 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
859 original change locked out all matches of empty strings. This has been
860 changed so that one match of an empty string per line is recognized.
861 Subsequent searches on the same line (for colouring or for --only-matching,
862 for example) do not recognize empty strings.
864 41. Applied a user patch to fix a number of spelling mistakes in comments.
866 42. Data lines longer than 65536 caused pcretest to crash.
868 43. Clarified the data type for length and startoffset arguments for pcre_exec
869 and pcre_dfa_exec in the function-specific man pages, where they were
870 explicitly stated to be in bytes, never having been updated. I also added
871 some clarification to the pcreapi man page.
873 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
874 a segmentation fault.
877 Version 8.32 30-November-2012
878 -----------------------------
880 1. Improved JIT compiler optimizations for first character search and single
881 character iterators.
883 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
884 Patch by Daniel Richard G.
886 3. Single character iterator optimizations in the JIT compiler.
888 4. Improved JIT compiler optimizations for character ranges.
890 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
891 Reported by Giuseppe D'Angelo.
893 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
894 of a line, was being set incorrectly in cases where .* appeared inside
895 atomic brackets at the start of a pattern, or where there was a subsequent
896 *PRUNE or *SKIP.
898 7. Improved instruction cache flush for POWER/PowerPC.
899 Patch by Daniel Richard G.
901 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
902 grep:
904 (a) There is now no limit to the number of patterns to be matched.
906 (b) An error is given if a pattern is too long.
908 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
909 are now supported.
911 (d) --exclude-from and --include-from (multiple use) have been added.
913 (e) Exclusions and inclusions now apply to all files and directories, not
914 just to those obtained from scanning a directory recursively.
916 (f) Multiple uses of -f and --file-list are now supported.
918 (g) In a Windows environment, the default for -d has been changed from
919 "read" (the GNU grep default) to "skip", because otherwise the presence
920 of a directory in the file list provokes an error.
922 (h) The documentation has been revised and clarified in places.
924 9. Improve the matching speed of capturing brackets.
926 10. Changed the meaning of \X so that it now matches a Unicode extended
927 grapheme cluster.
929 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
930 out POSIX threads when JIT support is configured.
932 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
934 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
935 structure could go wrong in environments where size_t is not the same size
936 as int.
938 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
939 set.
941 15. The EBCDIC support had decayed; later updates to the code had included
942 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
943 general tidy up of EBCDIC-related issues, and the documentation was also
944 not quite right. There is now a test that can be run on ASCII systems to
945 check some of the EBCDIC-related things (but is it not a full test).
947 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
948 in a small tidy to the code.
950 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
952 18. If the --only-matching (-o) option in pcregrep is specified multiple
953 times, each one causes appropriate output. For example, -o1 -o2 outputs the
954 substrings matched by the 1st and 2nd capturing parentheses. A separating
955 string can be specified by --om-separator (default empty).
957 19. Improving the first n character searches.
959 20. Turn case lists for horizontal and vertical white space into macros so that
960 they are defined only once.
962 21. This set of changes together give more compatible Unicode case-folding
963 behaviour for characters that have more than one other case when UCP
964 support is available.
966 (a) The Unicode property table now has offsets into a new table of sets of
967 three or more characters that are case-equivalent. The MultiStage2.py
968 script that generates these tables (the pcre_ucd.c file) now scans
969 CaseFolding.txt instead of UnicodeData.txt for character case
970 information.
972 (b) The code for adding characters or ranges of characters to a character
973 class has been abstracted into a generalized function that also handles
974 case-independence. In UTF-mode with UCP support, this uses the new data
975 to handle characters with more than one other case.
977 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
978 whose other case is greater than 256 are now correctly matched
979 caselessly. Previously, the high codepoint matched the low one, but not
980 vice versa.
982 (d) The processing of \h, \H, \v, and \ in character classes now makes use
983 of the new class addition function, using character lists defined as
984 macros alongside the case definitions of 20 above.
986 (e) Caseless back references now work with characters that have more than
987 one other case.
989 (f) General caseless matching of characters with more than one other case
990 is supported.
992 22. Unicode character properties were updated from Unicode 6.2.0
994 23. Improved CMake support under Windows. Patch by Daniel Richard G.
996 24. Add support for 32-bit character strings, and UTF-32
998 25. Major JIT compiler update (code refactoring and bugfixing).
999 Experimental Sparc 32 support is added.
1001 26. Applied a modified version of Daniel Richard G's patch to create
1002 pcre.h.generic and config.h.generic by "make" instead of in the
1003 PrepareRelease script.
1005 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1006 pcre_compile.c when checking for a zero character.
1008 28. Introducing a native interface for JIT. Through this interface, the compiled
1009 machine code can be directly executed. The purpose of this interface is to
1010 provide fast pattern matching, so several sanity checks are not performed.
1011 However, feature tests are still performed. The new interface provides
1012 1.4x speedup compared to the old one.
1014 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1015 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1016 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1018 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1019 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1020 the "old" RFC 2279). Instead, it ended up passing a negative length to
1021 pcre_exec().
1023 31. Add support for GCC's visibility feature to hide internal functions.
1025 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1026 "unknown -C option" after outputting 0 or 1.
1028 33. There is now support for generating a code coverage report for the test
1029 suite in environments where gcc is the compiler and lcov is installed. This
1030 is mainly for the benefit of the developers.
1032 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1033 unaddressable using valgrind annotations, allowing valgrind to detect
1034 invalid memory accesses. This is mainly for the benefit of the developers.
1036 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1038 26. Give configure error if --enable-cpp but no C++ compiler found.
1041 Version 8.31 06-July-2012
1042 -------------------------
1044 1. Fixing a wrong JIT test case and some compiler warnings.
1046 2. Removed a bashism from the RunTest script.
1048 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1049 to unsigned type, result still unsigned" that was given by an MS compiler
1050 on encountering the code "-sizeof(xxx)".
1052 4. Partial matching support is added to the JIT compiler.
1054 5. Fixed several bugs concerned with partial matching of items that consist
1055 of more than one character:
1057 (a) /^(..)\1/ did not partially match "aba" because checking references was
1058 done on an "all or nothing" basis. This also applied to repeated
1059 references.
1061 (b) \R did not give a hard partial match if \r was found at the end of the
1062 subject.
1064 (c) \X did not give a hard partial match after matching one or more
1065 characters at the end of the subject.
1067 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1068 a partial match for the string "\r".
1070 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1071 a partial match for a CR character at the end of the subject string.
1073 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1074 running pcretest, the text "(JIT)" added to the output whenever JIT is
1075 actually used to run the match.
1077 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1078 or /S+[+] with a digit between 1 and 7.
1080 8. OP_NOT now supports any UTF character not just single-byte ones.
1082 9. (*MARK) control verb is now supported by the JIT compiler.
1084 10. The command "./RunTest list" lists the available tests without actually
1085 running any of them. (Because I keep forgetting what they all are.)
1089 12. Applied a (slightly modified) user-supplied patch that improves performance
1090 when the heap is used for recursion (compiled with --disable-stack-for-
1091 recursion). Instead of malloc and free for each heap frame each time a
1092 logical recursion happens, frames are retained on a chain and re-used where
1093 possible. This sometimes gives as much as 30% improvement.
1095 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1096 call.
1098 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1100 15. It is now possible to link pcretest with libedit as an alternative to
1101 libreadline.
1103 16. (*COMMIT) control verb is now supported by the JIT compiler.
1105 17. The Unicode data tables have been updated to Unicode 6.1.0.
1107 18. Added --file-list option to pcregrep.
1109 19. Added binary file support to pcregrep, including the -a, --binary-files,
1110 -I, and --text options.
1112 20. The madvise function is renamed for posix_madvise for QNX compatibility
1113 reasons. Fixed by Giuseppe D'Angelo.
1115 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1117 22. Changed the option for creating symbolic links for 16-bit man pages from
1118 -s to -sf so that re-installing does not cause issues.
1120 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1122 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1123 matches in certain environments (the workspace was not being correctly
1124 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1125 some of the workspace data at the beginning of a restart.
1127 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1128 was not doing so when it should - probably a typo introduced by SVN 528
1129 (change 8.10/14).
1131 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1132 \w+ when the character tables indicated that \x{c4} was a word character.
1133 There were several related cases, all because the tests for doing a table
1134 lookup were testing for characters less than 127 instead of 255.
1136 27. If a pattern contains capturing parentheses that are not used in a match,
1137 their slots in the ovector are set to -1. For those that are higher than
1138 any matched groups, this happens at the end of processing. In the case when
1139 there were back references that the ovector was too small to contain
1140 (causing temporary malloc'd memory to be used during matching), and the
1141 highest capturing number was not used, memory off the end of the ovector
1142 was incorrectly being set to -1. (It was using the size of the temporary
1143 memory instead of the true size.)
1145 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1146 ovector size, it uses memory at the end of the block that it has got.
1148 29. Check for an overlong MARK name and give an error at compile time. The
1149 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1151 30. JIT compiler update.
1153 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1154 Rill for the patch.
1156 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1158 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1160 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1161 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1162 was enabled.
1164 35. Improve JIT code generation for greedy plus quantifier.
1166 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1167 1 to "aa" instead of to an empty string. The bug affected repeated groups
1168 that could potentially match an empty string.
1170 37. Optimizing single character iterators in JIT.
1172 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1173 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1174 codepoints that are too big for the mode are faulted, and in a UTF mode,
1175 disallowed codepoints are also faulted.
1177 39. If PCRE was compiled with UTF support, in three places in the DFA
1178 matcher there was code that should only have been obeyed in UTF mode, but
1179 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1180 processing when bytes with values greater than 127 were present. In 16-bit
1181 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1182 both cases the values are those that cannot be the first data item in a UTF
1183 character. The three items that might have provoked this were recursions,
1184 possessively repeated groups, and atomic groups.
1186 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1187 and pcregrep, because some OS require shared objects to be explicitly
1188 passed to ld, causing the link step to fail if they are not.
1190 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1191 mode, patterns that started with \h* or \R* might be incorrectly matched.
1194 Version 8.30 04-February-2012
1195 -----------------------------
1197 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1198 name is defined in ctype.h.
1200 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1201 only in quite long subpatterns.
1203 3. Removed the function pcre_info(), which has been obsolete and deprecated
1204 since it was replaced by pcre_fullinfo() in February 2000.
1206 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1207 match a (*MARK), and the match failed at the start of the subject, a
1208 reference to memory before the start of the subject could occur. This bug
1209 was introduced by fix 17 of release 8.21.
1211 5. A reference to an unset group with zero minimum repetition was giving
1212 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1213 /(another)?(\1?)test/ matched against "hello world test". This bug was
1214 introduced in release 8.13.
1216 6. Add support for 16-bit character strings (a large amount of work involving
1217 many changes and refactorings).
1219 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1220 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1221 from a file.
1223 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1224 rounding is not applied in this particular case).
1226 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1227 if they appear, or are escaped, in patterns.
1229 10. Get rid of a number of -Wunused-but-set-variable warnings.
1231 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1232 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1233 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1234 also returns the mark "x". This bug applied to capturing parentheses,
1235 non-capturing parentheses, and atomic parentheses. It also applied to some
1236 assertions.
1238 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1239 information out of configure.ac instead of relying on pcre.h.generic, which
1240 is not stored in the repository.
1242 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1243 -lreadline.
1245 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1247 15. Applied Graycode's patch to put the top-level frame on the stack rather
1248 than the heap when not using the stack for recursion. This gives a
1249 performance improvement in many cases when recursion is not deep.
1251 16. Experimental code added to "pcretest -C" to output the stack frame size.
1254 Version 8.21 12-Dec-2011
1255 ------------------------
1257 1. Updating the JIT compiler.
1259 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1260 are added as well.
1262 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1263 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1264 calling _pcre_jit_exec. Some extra comments are added.
1266 4. (*MARK) settings inside atomic groups that do not contain any capturing
1267 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1268 was introduced by change 18 for 8.20.
1270 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1271 ECMA-262 standard.
1273 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1274 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1275 This bug was probably introduced by change 9 of 8.13.
1277 7. While fixing 6 above, I noticed that a number of other items were being
1278 incorrectly rejected as "not fixed length". This arose partly because newer
1279 opcodes had not been added to the fixed-length checking code. I have (a)
1280 corrected the bug and added tests for these items, and (b) arranged for an
1281 error to occur if an unknown opcode is encountered while checking for fixed
1282 length instead of just assuming "not fixed length". The items that were
1283 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1284 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1285 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1287 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1288 being incorrectly compiled and would have given unpredicatble results.
1290 9. A possessively repeated subpattern with minimum repeat count greater than
1291 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1292 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1293 the first (A) could occur when it should not.
1295 10. Add a cast and remove a redundant test from the code.
1297 11. JIT should use pcre_malloc/pcre_free for allocation.
1299 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1300 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1301 is anything other than /usr, -L is still shown).
1303 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1305 14. Perl does not support \N without a following name in a [] class; PCRE now
1306 also gives an error.
1308 15. If a forward reference was repeated with an upper limit of around 2000,
1309 it caused the error "internal error: overran compiling workspace". The
1310 maximum number of forward references (including repeats) was limited by the
1311 internal workspace, and dependent on the LINK_SIZE. The code has been
1312 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1313 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1314 of around 200,000 forward references. While doing this, I also speeded up
1315 the filling in of repeated forward references.
1317 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1318 incorrectly expecting the subject to contain another "a" after the start.
1320 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1321 in the match, the SKIP should be ignored. This was not happening; instead
1322 the SKIP was being treated as NOMATCH. For patterns such as
1323 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1324 tested.
1326 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1327 now much more compatible with Perl, in particular in cases where the result
1328 is a non-match for a non-anchored pattern. For example, if
1329 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1330 "m", where previously it did not return a name. A side effect of this
1331 change is that for partial matches, the last encountered mark name is
1332 returned, as for non matches. A number of tests that were previously not
1333 Perl-compatible have been moved into the Perl-compatible test files. The
1334 refactoring has had the pleasing side effect of removing one argument from
1335 the match() function, thus reducing its stack requirements.
1337 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1338 subsequent uses of /S (without +) incorrectly behaved like /S+.
1340 21. Retrieve executable code size support for the JIT compiler and fixing
1341 some warnings.
1343 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1344 not work when the shorter character appeared right at the end of the
1345 subject string.
1347 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1348 systems.
1350 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1351 output it when the /M option is used in pcretest.
1353 25. The CheckMan script was not being included in the distribution. Also, added
1354 an explicit "perl" to run Perl scripts from the PrepareRelease script
1355 because this is reportedly needed in Windows.
1357 26. If study data was being save in a file and studying had not found a set of
1358 "starts with" bytes for the pattern, the data written to the file (though
1359 never used) was taken from uninitialized memory and so caused valgrind to
1360 complain.
1362 27. Updated RunTest.bat as provided by Sheri Pierce.
1364 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1366 29. Computation of memory usage for the table of capturing group names was
1367 giving an unnecessarily large value.
1370 Version 8.20 21-Oct-2011
1371 ------------------------
1373 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1374 a POSIX class. After further experiments with Perl, which convinced me that
1375 Perl has bugs and confusions, a closing square bracket is no longer allowed
1376 in a POSIX name. This bug also affected patterns with classes that started
1377 with full stops.
1379 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1380 captured substring, but while checking the failing first alternative,
1381 substring 1 is temporarily captured. If the output vector supplied to
1382 pcre_exec() was not big enough for this capture, the yield of the function
1383 was still zero ("insufficient space for captured substrings"). This cannot
1384 be totally fixed without adding another stack variable, which seems a lot
1385 of expense for a edge case. However, I have improved the situation in cases
1386 such as /(a)(b)x|abc/ matched against "abc", where the return code
1387 indicates that fewer than the maximum number of slots in the ovector have
1388 been set.
1390 3. Related to (2) above: when there are more back references in a pattern than
1391 slots in the output vector, pcre_exec() uses temporary memory during
1392 matching, and copies in the captures as far as possible afterwards. It was
1393 using the entire output vector, but this conflicts with the specification
1394 that only 2/3 is used for passing back captured substrings. Now it uses
1395 only the first 2/3, for compatibility. This is, of course, another edge
1396 case.
1398 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1399 main code base, and can be used by building with --enable-jit. When this is
1400 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1401 runtime --no-jit option is given.
1403 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1404 ovector, the return from the function was zero, implying that there were
1405 other matches that did not fit. The correct "exactly full" value is now
1406 returned.
1408 6. If a subpattern that was called recursively or as a subroutine contained
1409 (*PRUNE) or any other control that caused it to give a non-standard return,
1410 invalid errors such as "Error -26 (nested recursion at the same subject
1411 position)" or even infinite loops could occur.
1413 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1414 computing the minimum length on reaching *ACCEPT, and so ended up with the
1415 wrong value of 1 rather than 0. Further investigation indicates that
1416 computing a minimum subject length in the presence of *ACCEPT is difficult
1417 (think back references, subroutine calls), and so I have changed the code
1418 so that no minimum is registered for a pattern that contains *ACCEPT.
1420 8. If (*THEN) was present in the first (true) branch of a conditional group,
1421 it was not handled as intended. [But see 16 below.]
1423 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1424 Sheri Pierce.
1426 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1427 the first byte in a match must be "a".
1429 11. Change 17 for 8.13 increased the recursion depth for patterns like
1430 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1431 pattern contains any instances of (*THEN). If it does not, the old
1432 optimizations are restored. It would be nice to do this on a per-group
1433 basis, but at the moment that is not feasible.
1435 12. In some environments, the output of pcretest -C is CRLF terminated. This
1436 broke RunTest's code that checks for the link size. A single white space
1437 character after the value is now allowed for.
1439 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1440 For "fr", it uses the Windows-specific input and output files.
1442 14. If (*THEN) appeared in a group that was called recursively or as a
1443 subroutine, it did not work as intended. [But see next item.]
1445 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1446 pattern fragments (but not containing any | characters). If A and B are
1447 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1448 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1449 D. In other words, Perl considers parentheses that do not contain any |
1450 characters to be part of a surrounding alternative, whereas PCRE was
1451 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1452 differently. PCRE now behaves in the same way as Perl, except in the case
1453 of subroutine/recursion calls such as (?1) which have in any case always
1454 been different (but PCRE had them first :-).
1456 16. Related to 15 above: Perl does not treat the | in a conditional group as
1457 creating alternatives. Such a group is treated in the same way as an
1458 ordinary group without any | characters when processing (*THEN). PCRE has
1459 been changed to match Perl's behaviour.
1461 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1462 RunGrepTest script failed.
1464 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1465 inevitable for groups that contain captures, but it can lead to a lot of
1466 stack use in large patterns. The old behaviour has been restored for atomic
1467 groups that do not contain any capturing parentheses.
1469 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1470 suppress the check for a minimum subject length at run time. (If it was
1471 given to pcre_exec() or pcre_dfa_exec() it did work.)
1473 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1474 fail to work when decoding hex characters in data strings in EBCDIC
1475 environments.
1477 21. It appears that in at least one Mac OS environment, the isxdigit() function
1478 is implemented as a macro that evaluates to its argument more than once,
1479 contravening the C 90 Standard (I haven't checked a later standard). There
1480 was an instance in pcretest which caused it to go wrong when processing
1481 \x{...} escapes in subject strings. The has been rewritten to avoid using
1482 things like p++ in the argument of isxdigit().
1485 Version 8.13 16-Aug-2011
1486 ------------------------
1488 1. The Unicode data tables have been updated to Unicode 6.0.0.
1490 2. Two minor typos in pcre_internal.h have been fixed.
1492 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1493 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1494 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1496 4. There were a number of related bugs in the code for matching backrefences
1497 caselessly in UTF-8 mode when codes for the characters concerned were
1498 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1499 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1500 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1501 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1502 2-byte code at the end of the subject (it thought there wasn't enough data
1503 left).
1505 5. Comprehensive information about what went wrong is now returned by
1506 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1507 as the output vector has at least 2 elements. The offset of the start of
1508 the failing character and a reason code are placed in the vector.
1510 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1511 now returned is for the first byte of the failing character, instead of the
1512 last byte inspected. This is an incompatible change, but I hope it is small
1513 enough not to be a problem. It makes the returned offset consistent with
1514 pcre_exec() and pcre_dfa_exec().
1516 7. pcretest now gives a text phrase as well as the error number when
1517 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1518 failure, the offset and reason code are output.
1520 8. When \R was used with a maximizing quantifier it failed to skip backwards
1521 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1522 back over a single character (\n). This seems wrong (because it treated the
1523 two characters as a single entity when going forwards), conflicts with the
1524 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1525 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1526 has been changed.
1528 9. Some internal refactoring has changed the processing so that the handling
1529 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1530 time (the PCRE_DOTALL option was changed this way some time ago: version
1531 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1532 which was always a bit of a fudge. It also means that there is one less
1533 argument for the match() function, which reduces its stack requirements
1534 slightly. This change also fixes an incompatibility with Perl: the pattern
1535 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1537 10. More internal refactoring has drastically reduced the number of recursive
1538 calls to match() for possessively repeated groups such as (abc)++ when
1539 using pcre_exec().
1541 11. While implementing 10, a number of bugs in the handling of groups were
1542 discovered and fixed:
1544 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1545 (a|)*(?1) gave a compile-time internal error.
1546 ((a|)+)+ did not notice that the outer group could match an empty string.
1547 (^a|^)+ was not marked as anchored.
1548 (.*a|.*)+ was not marked as matching at start or after a newline.
1550 12. Yet more internal refactoring has removed another argument from the match()
1551 function. Special calls to this function are now indicated by setting a
1552 value in a variable in the "match data" data block.
1554 13. Be more explicit in pcre_study() instead of relying on "default" for
1555 opcodes that mean there is no starting character; this means that when new
1556 ones are added and accidentally left out of pcre_study(), testing should
1557 pick them up.
1559 14. The -s option of pcretest has been documented for ages as being an old
1560 synonym of -m (show memory usage). I have changed it to mean "force study
1561 for every regex", that is, assume /S for every regex. This is similar to -i
1562 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1563 using it. It makes it easier to run collections of tests with and without
1564 study enabled, and thereby test pcre_study() more easily. All the standard
1565 tests are now run with and without -s (but some patterns can be marked as
1566 "never study" - see 20 below).
1568 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1569 restoration of the capturing data to the outer values was not happening
1570 correctly.
1572 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1573 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1574 pattern had matched an empty string, and so incorrectly returned a no
1575 match.
1577 17. There was optimizing code for the last branch of non-capturing parentheses,
1578 and also for the obeyed branch of a conditional subexpression, which used
1579 tail recursion to cut down on stack usage. Unfortunately, now that there is
1580 the possibility of (*THEN) occurring in these branches, tail recursion is
1581 no longer possible because the return has to be checked for (*THEN). These
1582 two optimizations have therefore been removed. [But see 8.20/11 above.]
1584 18. If a pattern containing \R was studied, it was assumed that \R always
1585 matched two bytes, thus causing the minimum subject length to be
1586 incorrectly computed because \R can also match just one byte.
1588 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1589 was incorrectly computed.
1591 20. If /S is present twice on a test pattern in pcretest input, it now
1592 *disables* studying, thereby overriding the use of -s on the command line
1593 (see 14 above). This is necessary for one or two tests to keep the output
1594 identical in both cases.
1596 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1597 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1599 22. When an atomic group that contained a capturing parenthesis was
1600 successfully matched, but the branch in which it appeared failed, the
1601 capturing was not being forgotten if a higher numbered group was later
1602 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1603 group 1 to "a", when in fact it should be unset. This applied to multi-
1604 branched capturing and non-capturing groups, repeated or not, and also to
1605 positive assertions (capturing in negative assertions does not happen
1606 in PCRE) and also to nested atomic groups.
1608 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1609 subject after a captured substring, to make it easier to tell which of a
1610 number of identical substrings has been captured.
1612 24. The way atomic groups are processed by pcre_exec() has been changed so that
1613 if they are repeated, backtracking one repetition now resets captured
1614 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1615 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1616 "aaa". Previously, it would have been "a". As part of this code
1617 refactoring, the way recursive calls are handled has also been changed.
1619 25. If an assertion condition captured any substrings, they were not passed
1620 back unless some other capturing happened later. For example, if
1621 (?(?=(a))a) was matched against "a", no capturing was returned.
1623 26. When studying a pattern that contained subroutine calls or assertions,
1624 the code for finding the minimum length of a possible match was handling
1625 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1626 group 1 called group 2 while simultaneously a separate group 2 called group
1627 1). A stack overflow occurred in this case. I have fixed this by limiting
1628 the recursion depth to 10.
1630 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1631 Fortmann. This supports explicit test numbers on the command line, and has
1632 argument validation and error reporting.
1634 28. An instance of \X with an unlimited repeat could fail if at any point the
1635 first character it looked at was a mark character.
1637 29. Some minor code refactoring concerning Unicode properties and scripts
1638 should reduce the stack requirement of match() slightly.
1640 30. Added the '=' option to pcretest to check the setting of unused capturing
1641 slots at the end of the pattern, which are documented as being -1, but are
1642 not included in the return count.
1644 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1645 compiled something random. Now it gives a compile-time error (as does
1646 Perl).
1648 32. A *MARK encountered during the processing of a positive assertion is now
1649 recorded and passed back (compatible with Perl).
1651 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1652 had alternative anchored branches, the search for a second match in a line
1653 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1654 matched the line "0102" twice. The same bug affected patterns that started
1655 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1656 twice.
1658 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1659 does, and because of capturing effects, quantifying parenthesized
1660 assertions may at times be useful. Quantifiers are now allowed for
1661 parenthesized assertions.
1663 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1665 36. \g was being checked for fancy things in a character class, when it should
1666 just be a literal "g".
1668 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1669 appearance of a nested POSIX class supersedes an apparent external class.
1670 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1671 unescaped square brackets may also appear as part of class names. For
1672 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1673 more like Perl. (But see 8.20/1 above.)
1675 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1676 was because it thought it was \N{name}, which is not supported).
1678 39. Add minix to OS list not supporting the -S option in pcretest.
1680 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1681 cannot analyze patterns in sufficient detail to catch mutual recursions
1682 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1683 subgroup is called recursively as a subpattern for a second time at the
1684 same position in the subject string. In previous releases this might have
1685 been caught by the recursion limit, or it might have run out of stack.
1687 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1688 happen only once. PCRE was, however incorrectly giving a compile time error
1689 "recursive call could loop indefinitely" because it cannot analyze the
1690 pattern in sufficient detail. The compile time test no longer happens when
1691 PCRE is compiling a conditional subpattern, but actual runaway loops are
1692 now caught at runtime (see 40 above).
1694 42. It seems that Perl allows any characters other than a closing parenthesis
1695 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1696 has been changed to be the same.
1698 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1699 as not to get warnings when autogen.sh is called. Also changed
1700 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1702 44. To help people who use pcregrep to scan files containing exceedingly long
1703 lines, the following changes have been made:
1705 (a) The default value of the buffer size parameter has been increased from
1706 8K to 20K. (The actual buffer used is three times this size.)
1708 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1709 PCRE is built.
1711 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1712 to be set at run time.
1714 (d) Numerical values in pcregrep options can be followed by K or M, for
1715 example --buffer-size=50K.
1717 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1718 given and the return code is set to 2.
1720 45. Add a pointer to the latest mark to the callout data block.
1722 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1723 partial match of an empty string instead of no match. This was specific to
1724 the use of ".".
1726 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1727 complete match instead of a partial match. This bug was dependent on both
1728 the PCRE_UTF8 and PCRE_DOTALL options being set.
1730 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1731 starting byte set, because \b was not being ignored.
1734 Version 8.12 15-Jan-2011
1735 ------------------------
1737 1. Fixed some typos in the markup of the man pages, and wrote a script that
1738 checks for such things as part of the documentation building process.
1740 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1741 --match-limit and --recursion-limit options (added for 8.11). In
1742 particular, this made one of the standard tests fail. (The integer value
1743 went into the wrong half of a long int.)
1745 3. If the --colour option was given to pcregrep with -v (invert match), it
1746 did strange things, either producing crazy output, or crashing. It should,
1747 of course, ignore a request for colour when reporting lines that do not
1748 match.
1750 4. Another pcregrep bug caused similar problems if --colour was specified with
1751 -M (multiline) and the pattern match finished with a line ending.
1753 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1754 matched in multiline mode, the following line was shown as part of the
1755 match. This seems wrong, so I have changed it.
1757 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1758 the check for further matches in the same line (so they could be coloured)
1759 to overrun the end of the current line. If another match was found, it was
1760 incorrectly shown (and then shown again when found in the next line).
1762 7. If pcregrep was compiled under Windows, there was a reference to the
1763 function pcregrep_exit() before it was defined. I am assuming this was
1764 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1765 reported by a user. I've moved the definition above the reference.
1768 Version 8.11 10-Dec-2010
1769 ------------------------
1771 1. (*THEN) was not working properly if there were untried alternatives prior
1772 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1773 backtracked to try for "b" instead of moving to the next alternative branch
1774 at the same level (in this case, to look for "c"). The Perl documentation
1775 is clear that when (*THEN) is backtracked onto, it goes to the "next
1776 alternative in the innermost enclosing group".
1778 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1779 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1780 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1781 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1782 (*THEN).
1784 3. If \s appeared in a character class, it removed the VT character from
1785 the class, even if it had been included by some previous item, for example
1786 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1787 of \s, but is part of the POSIX "space" class.)
1789 4. A partial match never returns an empty string (because you can always
1790 match an empty string at the end of the subject); however the checking for
1791 an empty string was starting at the "start of match" point. This has been
1792 changed to the "earliest inspected character" point, because the returned
1793 data for a partial match starts at this character. This means that, for
1794 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1795 (previously it gave "no match").
1797 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1798 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1799 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1800 has an implication that the given string is incomplete (because a partial
1801 match is preferred over a full match). For this reason, these items now
1802 give a partial match in this situation. [Aside: previously, the one case
1803 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1804 match rather than a full match, which was wrong by the old rules, but is
1805 now correct.]
1807 6. There was a bug in the handling of #-introduced comments, recognized when
1808 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1809 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1810 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1811 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1812 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1813 places in pcre_compile().
1815 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1816 comments when looking ahead for named forward references to subpatterns,
1817 the only newline sequence it recognized was NL. It now handles newlines
1818 according to the set newline convention.
1820 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1821 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1822 cater for a lack of strerror(). These oversights have been fixed.
1824 9. Added --match-limit and --recursion-limit to pcregrep.
1826 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1828 11. When the -o option was used, pcregrep was setting a return code of 1, even
1829 when matches were found, and --line-buffered was not being honoured.
1831 12. Added an optional parentheses number to the -o and --only-matching options
1832 of pcregrep.
1834 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1835 can match an empty string. The code to do it in pcretest and pcredemo
1836 needed fixing:
1838 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1839 only one byte after an empty string match just before CRLF (this case
1840 just got forgotten; "any" and "anycrlf" were OK).
1842 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1843 mode when an empty string match preceded an ASCII character followed by
1844 a non-ASCII character. (The code for advancing by one character rather
1845 than one byte was nonsense.)
1847 (c) The pcredemo.c sample program did not have any code at all to handle
1848 the cases when CRLF is a valid newline sequence.
1850 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1851 as a starting offset was within the subject string. There is now a new
1852 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1853 negative or greater than the length of the string. In order to test this,
1854 pcretest is extended to allow the setting of negative starting offsets.
1856 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1857 starting offset points to the beginning of a UTF-8 character was
1858 unnecessarily clumsy. I tidied it up.
1860 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1861 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1863 17. Nobody had reported that the --include_dir option, which was added in
1864 release 7.7 should have been called --include-dir (hyphen, not underscore)
1865 for compatibility with GNU grep. I have changed it to --include-dir, but
1866 left --include_dir as an undocumented synonym, and the same for
1867 --exclude-dir, though that is not available in GNU grep, at least as of
1868 release 2.5.4.
1870 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1871 characters from a string of bytes) have been redefined so as not to use
1872 loops, in order to improve performance in some environments. At the same
1873 time, I abstracted some of the common code into auxiliary macros to save
1874 repetition (this should not affect the compiled code).
1876 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1877 compile-time error is now given if \c is not followed by an ASCII
1878 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1879 different, and any byte value is allowed.)
1881 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1882 START_OPTIMIZE option, which is now allowed at compile time - but just
1883 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1884 to pcregrep and other applications that have no direct access to PCRE
1885 options. The new /Y option in pcretest sets this option when calling
1886 pcre_compile().
1888 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1889 back references. Groups containing recursive back references were forced to
1890 be atomic by that change, but in the case of named groups, the amount of
1891 memory required was incorrectly computed, leading to "Failed: internal
1892 error: code overflow". This has been fixed.
1894 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1895 pcretest.c, to avoid build problems in some Borland environments.
1898 Version 8.10 25-Jun-2010
1899 ------------------------
1901 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1902 THEN.
1904 2. (*ACCEPT) was not working when inside an atomic group.
1906 3. Inside a character class, \B is treated as a literal by default, but
1907 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1908 causes the error). The code is unchanged, but I tidied the documentation.
1910 4. Inside a character class, PCRE always treated \R and \X as literals,
1911 whereas Perl faults them if its -w option is set. I have changed PCRE so
1912 that it faults them when PCRE_EXTRA is set.
1914 5. Added support for \N, which always matches any character other than
1915 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1917 6. When compiling pcregrep with newer versions of gcc which may have
1918 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1919 declared with attribute warn_unused_result" were given. Just casting the
1920 result to (void) does not stop the warnings; a more elaborate fudge is
1921 needed. I've used a macro to implement this.
1923 7. Minor change to pcretest.c to avoid a compiler warning.
1925 8. Added four artifical Unicode properties to help with an option to make
1926 \s etc use properties (see next item). The new properties are: Xan
1927 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1929 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1930 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1931 this option. Modified pcretest to add /W to test this facility. Added
1932 REG_UCP to make it available via the POSIX interface.
1934 10. Added --line-buffered to pcregrep.
1936 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1937 studied, and the match started with a letter with a code point greater than
1938 127 whose first byte was different to the first byte of the other case of
1939 the letter, the other case of this starting letter was not recognized
1940 (#976).
1942 12. If a pattern that was studied started with a repeated Unicode property
1943 test, for example, \p{Nd}+, there was the theoretical possibility of
1944 setting up an incorrect bitmap of starting bytes, but fortunately it could
1945 not have actually happened in practice until change 8 above was made (it
1946 added property types that matched character-matching opcodes).
1948 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1949 possible starting bytes for non-anchored patterns.
1951 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1952 \R, and also a number of cases that involve Unicode properties, both
1953 explicit and implicit when PCRE_UCP is set.
1955 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1956 input, it could crash or give wrong results if characters with values
1957 greater than 0xc0 were present in the subject string. (Detail: it assumed
1958 UTF-8 input when processing these items.)
1960 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1961 size_t is 64-bit (#991).
1963 17. Added a check for running out of memory when PCRE is compiled with
1964 --disable-stack-for-recursion (#990).
1966 18. If the last data line in a file for pcretest does not have a newline on
1967 the end, a newline was missing in the output.
1969 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1970 less than 128) in its various bitmaps. However, there is a facility for
1971 generating tables according to the current locale when PCRE is compiled. It
1972 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1973 characters, are recognized by isspace() and therefore were getting set in
1974 these tables, and indeed these tables seem to approximate to ISO 8859. This
1975 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1976 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1977 which of course cannot start UTF-8 characters. I have changed the code so
1978 that only real ASCII characters (less than 128) and the correct starting
1979 bytes for UTF-8 encodings are set for characters greater than 127 when in
1980 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1981 altogether.)
1983 20. Added the /T option to pcretest so as to be able to run tests with non-
1984 standard character tables, thus making it possible to include the tests
1985 used for 19 above in the standard set of tests.
1987 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1988 reference to a subpattern the other side of a comment that contains an
1989 opening parenthesis caused either an internal compiling error, or a
1990 reference to the wrong subpattern.
1993 Version 8.02 19-Mar-2010
1994 ------------------------
1996 1. The Unicode data tables have been updated to Unicode 5.2.0.
1998 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1999 configured.
2001 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2002 original author of that file, following a query about its status.
2004 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2005 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2007 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2008 quantifier applied to a forward-referencing subroutine call, could compile
2009 incorrect code or give the error "internal error: previously-checked
2010 referenced subpattern not found".
2012 6. Both MS Visual Studio and Symbian OS have problems with initializing
2013 variables to point to external functions. For these systems, therefore,
2014 pcre_malloc etc. are now initialized to local functions that call the
2015 relevant global functions.
2017 7. There were two entries missing in the vectors called coptable and poptable
2018 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2019 I've fixed the data, and added a kludgy way of testing at compile time that
2020 the lengths are correct (equal to the number of opcodes).
2022 8. Following on from 7, I added a similar kludge to check the length of the
2023 eint vector in pcreposix.c.
2025 9. Error texts for pcre_compile() are held as one long string to avoid too
2026 much relocation at load time. To find a text, the string is searched,
2027 counting zeros. There was no check for running off the end of the string,
2028 which could happen if a new error number was added without updating the
2029 string.
2031 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2033 11. \K was not working if it appeared in an atomic group or in a group that
2034 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2035 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2036 the assertion is positive, but not if it is negative.
2038 12. Change 11 fortuitously reduced the size of the stack frame used in the
2039 "match()" function of pcre_exec.c by one pointer. Forthcoming
2040 implementation of support for (*MARK) will need an extra pointer on the
2041 stack; I have reserved it now, so that the stack frame size does not
2042 decrease.
2044 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2045 item in branch that calls a recursion is a subroutine call - as in the
2046 second branch in the above example - was incorrectly given the compile-
2047 time error "recursive call could loop indefinitely" because pcre_compile()
2048 was not correctly checking the subroutine for matching a non-empty string.
2050 14. The checks for overrunning compiling workspace could trigger after an
2051 overrun had occurred. This is a "should never occur" error, but it can be
2052 triggered by pathological patterns such as hundreds of nested parentheses.
2053 The checks now trigger 100 bytes before the end of the workspace.
2055 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2058 Version 8.01 19-Jan-2010
2059 ------------------------
2061 1. If a pattern contained a conditional subpattern with only one branch (in
2062 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2063 computed the wrong minimum data length (which is of course zero for such
2064 subpatterns). This could cause incorrect "no match" results.
2066 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2067 the pattern is reset in the first branch, pcre_compile() failed with
2068 "internal error: code overflow at offset...". This happened only when
2069 the reset was to the original external option setting. (An optimization
2070 abstracts leading options settings into an external setting, which was the
2071 cause of this.)
2073 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2074 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2075 assertion pattern did not match (meaning that the assertion was true), it
2076 was incorrectly treated as false if the SKIP had been reached during the
2077 matching. This also applied to assertions used as conditions.
2079 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2080 assertion subpattern, including such a pattern used as a condition,
2081 unpredictable results occurred, instead of the error return
2084 5. The C++ GlobalReplace function was not working like Perl for the special
2085 situation when an empty string is matched. It now does the fancy magic
2086 stuff that is necessary.
2088 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2089 removed. (These were left over from very, very early versions of PCRE.)
2091 7. Some cosmetic changes to the code to make life easier when compiling it
2092 as part of something else:
2094 (a) Change DEBUG to PCRE_DEBUG.
2096 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2097 called "current" as "current_branch", to prevent a collision with the
2098 Linux macro when compiled as a kernel module.
2100 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2101 prevent a collision with the Linux macro when compiled as a kernel
2102 module.
2104 8. In pcre_compile() there are some checks for integer overflows that used to
2105 cast potentially large values to (double). This has been changed to that
2106 when building, a check for int64_t is made, and if it is found, it is used
2107 instead, thus avoiding the use of floating point arithmetic. (There is no
2108 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2109 double.
2111 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2112 2005 (difference between two addresses compared to an unsigned value).
2114 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2115 custom one, because of the following reported problem in Windows:
2117 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2118 under Win32.
2119 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2120 therefore missing the function definition.
2121 - The compiler thus generates a "C" signature for the test function.
2122 - The linker fails to find the "C" function.
2123 - PCRE fails to configure if asked to do so against libbz2.
2125 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2126 messages were output:
2128 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2129 rerunning libtoolize, to keep the correct libtool macros in-tree.
2130 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2132 I have done both of these things.
2134 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2135 most of the time, it *can* run out if it is given a pattern that contains a
2136 runaway infinite recursion. I updated the discussion in the pcrestack man
2137 page.
2139 13. Now that we have gone to the x.xx style of version numbers, the minor
2140 version may start with zero. Using 08 or 09 is a bad idea because users
2141 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2142 interpreted as invalid octal numbers. I've updated the previous comment in
2143 configure.ac, and also added a check that gives an error if 08 or 09 are
2144 used.
2146 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2147 causing partial matching to fail when the end of the subject matched \W
2148 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2150 15. There were some discrepancies between the declarations in pcre_internal.h
2151 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2152 their definitions. The declarations used "const uschar *" and the
2153 definitions used USPTR. Even though USPTR is normally defined as "const
2154 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2155 reported that: "This difference in casting confuses some C++ compilers, for
2156 example, SunCC recognizes above declarations as different functions and
2157 generates broken code for hbpcre." I have changed the declarations to use
2158 USPTR.
2160 16. GNU libtool is named differently on some systems. The autogen.sh script now
2161 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2162 (FreeBSD).
2164 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2165 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2166 comment: "Figure out how to create a longlong from a string: strtoll and
2167 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2168 instance, but it only takes 2 args instead of 3!"
2170 18. A subtle bug concerned with back references has been fixed by a change of
2171 specification, with a corresponding code fix. A pattern such as
2172 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2173 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2174 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2175 same bug. Such groups have to be quantified to be useful, or contained
2176 inside another quantified group. (If there's no repetition, the reference
2177 can never match.) The problem arises because, having left the group and
2178 moved on to the rest of the pattern, a later failure that backtracks into
2179 the group uses the captured value from the final iteration of the group
2180 rather than the correct earlier one. I have fixed this in PCRE by forcing
2181 any group that contains a reference to itself to be an atomic group; that
2182 is, there cannot be any backtracking into it once it has completed. This is
2183 similar to recursive and subroutine calls.
2186 Version 8.00 19-Oct-09
2187 ----------------------
2189 1. The table for translating pcre_compile() error codes into POSIX error codes
2190 was out-of-date, and there was no check on the pcre_compile() error code
2191 being within the table. This could lead to an OK return being given in
2192 error.
2194 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2195 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2196 in a Windows environment.
2198 3. The pcregrep --count option prints the count for each file even when it is
2199 zero, as does GNU grep. However, pcregrep was also printing all files when
2200 --files-with-matches was added. Now, when both options are given, it prints
2201 counts only for those files that have at least one match. (GNU grep just
2202 prints the file name in this circumstance, but including the count seems
2203 more useful - otherwise, why use --count?) Also ensured that the
2204 combination -clh just lists non-zero counts, with no names.
2206 4. The long form of the pcregrep -F option was incorrectly implemented as
2207 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2208 but it seems right to fix it, and I didn't think it was worth preserving
2209 the old behaviour.
2211 5. The command line items --regex=pattern and --regexp=pattern were not
2212 recognized by pcregrep, which required --regex pattern or --regexp pattern
2213 (with a space rather than an '='). The man page documented the '=' forms,
2214 which are compatible with GNU grep; these now work.
2216 6. No libpcreposix.pc file was created for pkg-config; there was just
2217 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2219 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2220 when UCP support is not needed, by modifying the Python script that
2221 generates it from Unicode data files. This should not matter if the module
2222 is correctly used as a library, but I received one complaint about 50K of
2223 unwanted data. My guess is that the person linked everything into his
2224 program rather than using a library. Anyway, it does no harm.
2226 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2227 was a minimum greater than 1 for a wide character in a possessive
2228 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2229 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2230 character. Chaos in the form of incorrect output or a compiling loop could
2231 result.
2233 9. The restrictions on what a pattern can contain when partial matching is
2234 requested for pcre_exec() have been removed. All patterns can now be
2235 partially matched by this function. In addition, if there are at least two
2236 slots in the offset vector, the offset of the earliest inspected character
2237 for the match and the offset of the end of the subject are set in them when
2238 PCRE_ERROR_PARTIAL is returned.
2240 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2241 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2242 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2243 and may be more useful for multi-segment matching.
2245 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2246 used to be given if ever the end of the subject was reached; now it is
2247 given only if matching could not proceed because another character was
2248 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2249 string "Z", which now yields "no match" instead of "partial match". In the
2250 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2251 final character ended with (*FAIL).
2253 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2254 if the pattern had a "must contain" character that was already found in the
2255 earlier partial match, unless partial matching was again requested. For
2256 example, with the pattern /dog.(body)?/, the "must contain" character is
2257 "g". If the first part-match was for the string "dog", restarting with
2258 "sbody" failed. This bug has been fixed.
2260 13. The string returned by pcre_dfa_exec() after a partial match has been
2261 changed so that it starts at the first inspected character rather than the
2262 first character of the match. This makes a difference only if the pattern
2263 starts with a lookbehind assertion or \b or \B (\K is not supported by
2264 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2265 matching functions compatible, and I think it's the right thing to do.
2267 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2268 so that the demonstration program is easily available in environments where
2269 PCRE has not been installed from source.
2271 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2272 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2273 library.
2275 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2276 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2277 is not the first non-POSIX option to be added. Clearly some people find
2278 these options useful.
2280 17. If a caller to the POSIX matching function regexec() passes a non-zero
2281 value for nmatch with a NULL value for pmatch, the value of
2282 nmatch is forced to zero.
2284 18. RunGrepTest did not have a test for the availability of the -u option of
2285 the diff command, as RunTest does. It now checks in the same way as
2286 RunTest, and also checks for the -b option.
2288 19. If an odd number of negated classes containing just a single character
2289 interposed, within parentheses, between a forward reference to a named
2290 subpattern and the definition of the subpattern, compilation crashed with
2291 an internal error, complaining that it could not find the referenced
2292 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2293 [The bug was that it was starting one character too far in when skipping
2294 over the character class, thus treating the ] as data rather than
2295 terminating the class. This meant it could skip too much.]
2297 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2298 /g option in pcretest when the pattern contains \K, which makes it possible
2299 to have an empty string match not at the start, even when the pattern is
2300 anchored. Updated pcretest and pcredemo to use this option.
2302 21. If the maximum number of capturing subpatterns in a recursion was greater
2303 than the maximum at the outer level, the higher number was returned, but
2304 with unset values at the outer level. The correct (outer level) value is
2305 now given.
2307 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2308 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2309 make it do so. The string so far is captured, making this feature
2310 compatible with Perl.
2312 23. The tests have been re-organized, adding tests 11 and 12, to make it
2313 possible to check the Perl 5.10 features against Perl 5.10.
2315 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2316 pattern matches a fixed length string. PCRE did not allow this; now it
2317 does. Neither allows recursion.
2319 25. I finally figured out how to implement a request to provide the minimum
2320 length of subject string that was needed in order to match a given pattern.
2321 (It was back references and recursion that I had previously got hung up
2322 on.) This code has now been added to pcre_study(); it finds a lower bound
2323 to the length of subject needed. It is not necessarily the greatest lower
2324 bound, but using it to avoid searching strings that are too short does give
2325 some useful speed-ups. The value is available to calling programs via
2326 pcre_fullinfo().
2328 26. While implementing 25, I discovered to my embarrassment that pcretest had
2329 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2330 study optimizations had never been tested with that matching function.
2331 Oops. What is worse, even when it was passed study data, there was a bug in
2332 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2333 were also very few tests of studied patterns with pcre_dfa_exec().
2335 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2336 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2337 on the other side of the coin, they are no longer allowed to have different
2338 names, because these cannot be distinguished in PCRE, and this has caused
2339 confusion. (This is a difference from Perl.)
2341 28. When duplicate subpattern names are present (necessarily with different
2342 numbers, as required by 27 above), and a test is made by name in a
2343 conditional pattern, either for a subpattern having been matched, or for
2344 recursion in such a pattern, all the associated numbered subpatterns are
2345 tested, and the overall condition is true if the condition is true for any
2346 one of them. This is the way Perl works, and is also more like the way
2347 testing by number works.
2350 Version 7.9 11-Apr-09
2351 ---------------------
2353 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2354 (pcretest), all targets were linked against these libraries. This included
2355 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2356 libraries. This caused unwanted dependencies to be created. This problem
2357 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2358 pcretest is linked with readline.
2360 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2361 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2362 moved outside it again, because FALSE and TRUE are already defined in AIX,
2363 but BOOL is not.
2365 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2366 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2368 4. The pcregrep documentation said spaces were inserted as well as colons (or
2369 hyphens) following file names and line numbers when outputting matching
2370 lines. This is not true; no spaces are inserted. I have also clarified the
2371 wording for the --colour (or --color) option.
2373 5. In pcregrep, when --colour was used with -o, the list of matching strings
2374 was not coloured; this is different to GNU grep, so I have changed it to be
2375 the same.
2377 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2378 each matching line was coloured. Now it goes on to look for further matches
2379 of any of the test patterns, which is the same behaviour as GNU grep.
2381 7. A pattern that could match an empty string could cause pcregrep to loop; it
2382 doesn't make sense to accept an empty string match in pcregrep, so I have
2383 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2384 seems to be how GNU grep behaves. [But see later change 40 for release
2385 8.33.]
2387 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2388 start or after a newline", because the conditional assertion was not being
2389 correctly handled. The rule now is that both the assertion and what follows
2390 in the first alternative must satisfy the test.
2392 9. If auto-callout was enabled in a pattern with a conditional group whose
2393 condition was an assertion, PCRE could crash during matching, both with
2394 pcre_exec() and pcre_dfa_exec().
2396 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2397 used for matching.
2399 11. Unicode property support in character classes was not working for
2400 characters (bytes) greater than 127 when not in UTF-8 mode.
2402 12. Added the -M command line option to pcretest.
2404 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2406 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2408 16. Added comments and documentation about mis-use of no_arg in the C++
2409 wrapper.
2411 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2412 from Martin Jerabek that uses macro names for all relevant character and
2413 string constants.
2415 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2416 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2417 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2418 these, but not everybody uses configure.
2420 19. A conditional group that had only one branch was not being correctly
2421 recognized as an item that could match an empty string. This meant that an
2422 enclosing group might also not be so recognized, causing infinite looping
2423 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2424 with the subject "ab", where knowledge that the repeated group can match
2425 nothing is needed in order to break the loop.
2427 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2428 exec(), but without supplying a callout function, matching went wrong.
2430 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2431 leak if the size of the offset vector was greater than 30. When the vector
2432 is smaller, the saved offsets during recursion go onto a local stack
2433 vector, but for larger vectors malloc() is used. It was failing to free
2434 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2435 error, in fact).
2437 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2438 heapframe that is used only when UTF-8 support is enabled. This caused no
2439 problem, but was untidy.
2441 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2442 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2443 included within another project.
2445 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2446 slightly modified by me:
2448 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2449 not building pcregrep.
2451 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2452 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2454 25. Forward references, both numeric and by name, in patterns that made use of
2455 duplicate group numbers, could behave incorrectly or give incorrect errors,
2456 because when scanning forward to find the reference group, PCRE was not
2457 taking into account the duplicate group numbers. A pattern such as
2458 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2460 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2461 the feature of a custom pointer more persuasive (as requested by a user).
2463 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2464 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2465 pre-defined.
2467 28. Added support for (*UTF8) at the start of a pattern.
2469 29. Arrange for flags added by the "release type" setting in CMake to be shown
2470 in the configuration summary.
2473 Version 7.8 05-Sep-08
2474 ---------------------
2476 1. Replaced UCP searching code with optimized version as implemented for Ad
2477 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2478 stage table and inline lookup instead of a function, giving speed ups of 2
2479 to 5 times on some simple patterns that I tested. Permission was given to
2480 distribute the MultiStage2.py script that generates the tables (it's not in
2481 the tarball, but is in the Subversion repository).
2483 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2484 scripts.
2486 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2487 a group with a zero qualifier. The result of the study could be incorrect,
2488 or the function might crash, depending on the pattern.
2490 4. Caseless matching was not working for non-ASCII characters in back
2491 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2492 It now works when Unicode Property Support is available.
2494 5. In pcretest, an escape such as \x{de} in the data was always generating
2495 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2496 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2497 truncation.
2499 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2501 7. Added two (int) casts to pcregrep when printing the difference of two
2502 pointers, in case they are 64-bit values.
2504 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2505 test 2 if it fails.
2507 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2508 and a #define of that name to empty if it is not externally set. This is to
2509 allow users of MSVC to set it if necessary.
2511 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2512 the convenience functions in the pcre_get.c source file.
2514 11. An option change at the start of a pattern that had top-level alternatives
2515 could cause overwriting and/or a crash. This command provoked a crash in
2516 some environments:
2518 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2520 This potential security problem was recorded as CVE-2008-2371.
2522 12. For a pattern where the match had to start at the beginning or immediately
2523 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2524 pcre_dfa_exec() could read past the end of the passed subject if there was
2525 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2526 pcretest so that it places the subject at the end of its malloc-ed buffer.
2528 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2529 exec() might read past the end of the data buffer in UTF-8 mode.
2531 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2532 the data contained the byte 0x85 as part of a UTF-8 character within its
2533 first line. This applied both to normal and DFA matching.
2535 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2536 /^[^d]*?$/8 failed to match "abc".
2538 16. Added a missing copyright notice to pcrecpp_internal.h.
2540 17. Make it more clear in the documentation that values returned from
2541 pcre_exec() in ovector are byte offsets, not character counts.
2543 18. Tidied a few places to stop certain compilers from issuing warnings.
2545 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2546 supplied by Stefan Weber. I made a further small update for 7.8 because
2547 there is a change of source arrangements: the pcre_searchfuncs.c module is
2548 replaced by pcre_ucd.c.
2551 Version 7.7 07-May-08
2552 ---------------------
2554 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2555 a string to a long long, pretend we don't even have a long long." This is
2556 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2558 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2559 pre-7.6 versions, which defined a global no_arg variable instead of putting
2560 it in the RE class. (See also #8 below.)
2562 3. Remove a line of dead code, identified by coverity and reported by Nuno
2563 Lopes.
2565 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2567 (1) The include/exclude patterns were being applied to the whole pathnames
2568 of files, instead of just to the final components.
2570 (2) If there was more than one level of directory, the subdirectories were
2571 skipped unless they satisfied the include/exclude conditions. This is
2572 inconsistent with GNU grep (and could even be seen as contrary to the
2573 pcregrep specification - which I improved to make it absolutely clear).
2574 The action now is always to scan all levels of directory, and just
2575 apply the include/exclude patterns to regular files.
2577 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2578 --exclude_dir in the tests to avoid scanning .svn directories.
2580 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2581 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2582 doesn't support NULs in patterns.
2584 7. Added some missing "const"s to declarations of static tables in
2585 pcre_compile.c and pcre_dfa_exec.c.
2587 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2588 caused by fix #2 above. (Subsequently also a second patch to fix the
2589 first patch. And a third patch - this was a messy problem.)
2591 9. Applied Craig's patch to remove the use of push_back().
2593 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2594 matching function regexec().
2596 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2597 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2598 references. PCRE supports relative numbers with this syntax (I don't think
2599 Oniguruma does).
2601 12. Previously, a group with a zero repeat such as (...){0} was completely
2602 omitted from the compiled regex. However, this means that if the group
2603 was called as a subroutine from elsewhere in the pattern, things went wrong
2604 (an internal error was given). Such groups are now left in the compiled
2605 pattern, with a new opcode that causes them to be skipped at execution
2606 time.
2608 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2609 to the way PCRE behaves:
2611 (a) A lone ] character is dis-allowed (Perl treats it as data).
2613 (b) A back reference to an unmatched subpattern matches an empty string
2614 (Perl fails the current match path).
2616 (c) A data ] in a character class must be notated as \] because if the
2617 first data character in a class is ], it defines an empty class. (In
2618 Perl it is not possible to have an empty class.) The empty class []
2619 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2620 The negative empty class [^] matches any one character, independently
2621 of the DOTALL setting.
2623 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2624 non-existent subpattern following a character class starting with ']' and
2625 containing () gave an internal compiling error instead of "reference to
2626 non-existent subpattern". Fortunately, when the pattern did exist, the
2627 compiled code was correct. (When scanning forwards to check for the
2628 existence of the subpattern, it was treating the data ']' as terminating
2629 the class, so got the count wrong. When actually compiling, the reference
2630 was subsequently set up correctly.)
2632 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2633 it was being rejected as not supported by pcre_dfa_exec(), even though
2634 other assertions are supported. I have made pcre_dfa_exec() support
2635 (*FAIL).
2637 16. The implementation of 13c above involved the invention of a new opcode,
2638 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2639 cannot be changed at match time, I realized I could make a small
2640 improvement to matching performance by compiling OP_ALLANY instead of
2641 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2642 on the OP_ANY path.
2644 17. Compiling pcretest on Windows with readline support failed without the
2645 following two fixes: (1) Make the unistd.h include conditional on
2646 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2648 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2649 ncurses library to be included for pcretest when ReadLine support is
2650 requested, but also to allow for it to be overridden. This patch came from
2651 Daniel Bergström.
2653 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2654 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2655 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2656 spotting this.
2659 Version 7.6 28-Jan-08
2660 ---------------------
2662 1. A character class containing a very large number of characters with
2663 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2664 overflow.
2666 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2667 HAVE_LONG_LONG is not defined.
2669 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2670 bring them up to date and include new features. This patch includes:
2672 - Fixed PH's badly added libz and libbz2 support.
2673 - Fixed a problem with static linking.
2674 - Added pcredemo. [But later removed - see 7 below.]
2675 - Fixed dftables problem and added an option.
2676 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2678 - Added readline support for pcretest.
2679 - Added an listing of the option settings after cmake has run.
2681 4. A user submitted a patch to Makefile that makes it easy to create
2682 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2683 Makefile.am that cause it to include this special target, without
2684 affecting anything else. Note that the same mingw target plus all
2685 the other distribution libraries and programs are now supported
2686 when configuring with CMake (see 6 below) instead of with
2687 Configure/Make.
2689 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2690 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2691 exported in the Windows port". It has not yet been confirmed that the patch
2692 solves the problem, but it does no harm.
2694 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2695 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2696 with CMake, and also correct the comment about stack recursion.
2698 7. Remove the automatic building of pcredemo from the ./configure system and
2699 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2700 of a program that users should build themselves after PCRE is installed, so
2701 building it automatically is not really right. What is more, it gave
2702 trouble in some build environments.
2704 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2707 Version 7.5 10-Jan-08
2708 ---------------------
2710 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2711 values in parens when parsing an RE using the C++ wrapper."
2713 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2714 Characters greater than 255 were excluded from the class instead of being
2715 included.
2717 3. The same bug as (2) above applied to negated POSIX classes such as
2718 [:^space:].
2720 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2721 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2722 I have changed it.
2724 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2725 first named subpattern) and a construct such as (?&a) would reference the
2726 first named subpattern whose name started with "a" (in other words, the
2727 length check was missing). Both these problems are fixed. "Subpattern name
2728 expected" is now given for (?&) (a zero-length name), and this patch also
2729 makes it give the same error for \k'' (previously it complained that that
2730 was a reference to a non-existent subpattern).
2732 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2733 this is right because (?- can be followed by option settings as well as by
2734 digits. I have, however, made the messages clearer.
2736 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2737 than the number used in the conditional) now cause a compile-time error.
2738 This is actually not compatible with Perl, which accepts such patterns, but
2739 treats the conditional as always being FALSE (as PCRE used to), but it
2740 seems to me that giving a diagnostic is better.
2742 8. Change "alphameric" to the more common word "alphanumeric" in comments
2743 and messages.
2745 9. Fix two occurrences of "backslash" in comments that should have been
2746 "backspace".
2748 10. Remove two redundant lines of code that can never be obeyed (their function
2749 was moved elsewhere).
2751 11. The program that makes PCRE's Unicode character property table had a bug
2752 which caused it to generate incorrect table entries for sequences of
2753 characters that have the same character type, but are in different scripts.
2754 It amalgamated them into a single range, with the script of the first of
2755 them. In other words, some characters were in the wrong script. There were
2756 thirteen such cases, affecting characters in the following ranges:
2758 U+002b0 - U+002c1
2759 U+0060c - U+0060d
2760 U+0061e - U+00612
2761 U+0064b - U+0065e
2762 U+0074d - U+0076d
2763 U+01800 - U+01805
2764 U+01d00 - U+01d77
2765 U+01d9b - U+01dbf
2766 U+0200b - U+0200f
2767 U+030fc - U+030fe
2768 U+03260 - U+0327f
2769 U+0fb46 - U+0fbb1
2770 U+10450 - U+1049d
2772 12. The -o option (show only the matching part of a line) for pcregrep was not
2773 compatible with GNU grep in that, if there was more than one match in a
2774 line, it showed only the first of them. It now behaves in the same way as
2775 GNU grep.
2777 13. If the -o and -v options were combined for pcregrep, it printed a blank
2778 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2779 does the same. The return code can be used to tell if there were any
2780 non-matching lines.
2782 14. Added --file-offsets and --line-offsets to pcregrep.
2784 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2785 infinitely looping recursion. The bug was that positive lookaheads were not
2786 being skipped when checking for a possible empty match (negative lookaheads
2787 and both kinds of lookbehind were skipped).
2789 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2790 inclusion of <windows.h> to before rather than after the definition of
2791 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2793 17. Specifying a possessive quantifier with a specific limit for a Unicode
2794 character property caused pcre_compile() to compile bad code, which led at
2795 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2796 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2797 caused the error; without that there was no problem.
2799 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2801 19. Added --enable-pcretest-libreadline.
2803 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2804 RE::GlobalReplace(). As a result, the number of replacements returned was
2805 double what it should be. I removed one of the increments, but Craig sent a
2806 later patch that removed the other one (the right fix) and added unit tests
2807 that check the return values (which was not done before).
2809 21. Several CMake things:
2811 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2812 the names libpcre and libpcreposix, not just pcre and pcreposix.
2814 (2) The above change means that pcretest and pcregrep are now correctly
2815 linked with the newly-built libraries, not previously installed ones.
2819 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2820 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2821 UTF-8 newline character). The key issue is that the pattern starts .*;
2822 this means that the match must be either at the beginning, or after a
2823 newline. The bug was in the code for advancing after a failed match and
2824 checking that the new position followed a newline. It was not taking
2825 account of UTF-8 characters correctly.
2827 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2828 character classes. PCRE was not treating the sequence [:...:] as a
2829 character class unless the ... were all letters. Perl, however, seems to
2830 allow any characters between [: and :], though of course it rejects as
2831 unknown any "names" that contain non-letters, because all the known class
2832 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2833 for example, whereas PCRE did not - it did not recognize a POSIX character
2834 class. This seemed a bit dangerous, so the code has been changed to be
2835 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2836 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2837 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2838 Perl does, and where it didn't before.
2840 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2841 Windows environments %n is disabled by default.
2844 Version 7.4 21-Sep-07
2845 ---------------------
2847 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2848 means that a class such as [\s] counted as "explicit reference to CR or
2849 LF". That isn't really right - the whole point of the change was to try to
2850 help when there was an actual mention of one of the two characters. So now
2851 the change happens only if \r or \n (or a literal CR or LF) character is
2852 encountered.
2854 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2855 of both had grown to the point where there were only 3 bits left.
2856 Fortunately, there was spare space in the data structure, and so I have
2857 moved the internal flags into a new 16-bit field to free up more option
2858 bits.
2860 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2861 but did not set the internal JCHANGED flag - either of these is enough to
2862 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2863 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2864 start sets both bits.
2866 4. Added options (at build time, compile time, exec time) to change \R from
2867 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2869 5. doc/pcresyntax.html was missing from the distribution.
2871 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2872 compatibility, even though it is no longer used.
2874 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2875 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2876 windows.h file is present (where different names are used). [This was
2877 reversed later after testing - see 16 below.]
2879 8. Changed all #include <config.h> to #include "config.h". There were also
2880 some further <pcre.h> cases that I changed to "pcre.h".
2882 9. When pcregrep was used with the --colour option, it missed the line ending
2883 sequence off the lines that it output.
2885 10. It was pointed out to me that arrays of string pointers cause lots of
2886 relocations when a shared library is dynamically loaded. A technique of
2887 using a single long string with a table of offsets can drastically reduce
2888 these. I have refactored PCRE in four places to do this. The result is
2889 dramatic:
2891 Originally: 290
2892 After changing UCP table: 187
2893 After changing error message table: 43
2894 After changing table of "verbs" 36
2895 After changing table of Posix names 22
2897 Thanks to the folks working on Gregex for glib for this insight.
2899 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2900 unicode-properties was also set.
2902 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2904 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2905 checked only for CRLF.
2907 14. Added casts to pcretest.c to avoid compiler warnings.
2909 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2911 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2912 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2913 entirely. This removes changes made in 7 above.
2915 17. The CMake files have been updated, and there is now more information about
2916 building with CMake in the NON-UNIX-USE document.
2919 Version 7.3 28-Aug-07
2920 ---------------------
2922 1. In the rejigging of the build system that eventually resulted in 7.1, the
2923 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2924 brackets there is not right, since it causes compilers to look for an
2925 installed pcre.h, not the version that is in the source that is being
2926 compiled (which of course may be different). I have changed it back to:
2928 #include "pcre.h"
2930 I have a vague recollection that the change was concerned with compiling in
2931 different directories, but in the new build system, that is taken care of
2932 by the VPATH setting the Makefile.
2934 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2935 when the subject happened to end in the byte 0x85 (e.g. if the last
2936 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2937 characters but of course it shouldn't be taken as a newline when it is part
2938 of another character. The bug was that, for an unlimited repeat of . in
2939 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2940 characters when looking for a newline.
2942 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2944 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2945 in debug output.
2947 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2948 long printing in the pcrecpp unittest when running under MinGW.
2950 6. ESC_K was left out of the EBCDIC table.
2952 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2953 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2954 limit also applies to "virtual nesting" when a pattern is recursive, and in
2955 this case 1000 isn't so big. I have been able to remove this limit at the
2956 expense of backing off one optimization in certain circumstances. Normally,
2957 when pcre_exec() would call its internal match() function recursively and
2958 immediately return the result unconditionally, it uses a "tail recursion"
2959 feature to save stack. However, when a subpattern that can match an empty
2960 string has an unlimited repetition quantifier, it no longer makes this
2961 optimization. That gives it a stack frame in which to save the data for
2962 checking that an empty string has been matched. Previously this was taken
2963 from the 1000-entry workspace that had been reserved. So now there is no
2964 explicit limit, but more stack is used.
2966 8. Applied Daniel's patches to solve problems with the import/export magic
2967 syntax that is required for Windows, and which was going wrong for the
2968 pcreposix and pcrecpp parts of the library. These were overlooked when this
2969 problem was solved for the main library.
2971 9. There were some crude static tests to avoid integer overflow when computing
2972 the size of patterns that contain repeated groups with explicit upper
2973 limits. As the maximum quantifier is 65535, the maximum group length was
2974 set at 30,000 so that the product of these two numbers did not overflow a
2975 32-bit integer. However, it turns out that people want to use groups that
2976 are longer than 30,000 bytes (though not repeat them that many times).
2977 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2978 made it possible to implement the integer overflow checks in a much more
2979 dynamic way, which I have now done. The artificial limitation on group
2980 length has been removed - we now have only the limit on the total length of
2981 the compiled pattern, which depends on the LINK_SIZE setting.
2983 10. Fixed a bug in the documentation for get/copy named substring when
2984 duplicate names are permitted. If none of the named substrings are set, the
2985 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2986 empty string.
2988 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2989 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2990 because the ] is interpreted as the first data character and the
2991 terminating ] is not found. PCRE has been made compatible with Perl in this
2992 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2993 cause memory overwriting.
2995 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2996 string has been matched (to stop an infinite loop). It was not recognizing
2997 a conditional subpattern that could match an empty string if that
2998 subpattern was within another subpattern. For example, it looped when
2999 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3000 condition was not nested. This bug has been fixed.
3002 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3003 past the start of the subject in the presence of bytes with the top bit
3004 set, for example "\x8aBCD".
3006 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3007 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3009 14. Optimized (?!) to (*FAIL).
3011 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3012 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3013 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3014 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3015 does: it's just the validity check that is more restrictive.
3017 16. Inserted checks for integer overflows during escape sequence (backslash)
3018 processing, and also fixed erroneous offset values for syntax errors during
3019 backslash processing.
3021 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3022 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3024 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3025 caused an overrun.
3027 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3028 something other than just ASCII characters) inside a group that had an
3029 unlimited repeat caused a loop at compile time (while checking to see
3030 whether the group could match an empty string).
3032 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3033 [\P{Any}] did so. (Error in the code for printing property names.)
3035 21. An orphan \E inside a character class could cause a crash.
3037 22. A repeated capturing bracket such as (A)? could cause a wild memory
3038 reference during compilation.
3040 23. There are several functions in pcre_compile() that scan along a compiled
3041 expression for various reasons (e.g. to see if it's fixed length for look
3042 behind). There were bugs in these functions when a repeated \p or \P was
3043 present in the pattern. These operators have additional parameters compared
3044 with \d, etc, and these were not being taken into account when moving along
3045 the compiled data. Specifically:
3047 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3048 length.
3050 (b) An item such as \pL+ within a repeated group could cause crashes or
3051 loops.
3053 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3054 "reference to non-existent subpattern" error.
3056 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3058 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3059 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3061 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3063 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3064 character were causing crashes (broken optimization).
3066 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3067 \p or \P) caused a compile-time loop.
3069 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3070 break. For example, the unstudied pattern [\r\n]A does not match the string
3071 "\r\nA" because change 7.0/46 below moves the current point on by two
3072 characters after failing to match at the start. However, the pattern \nA
3073 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3074 the same is true. There doesn't seem any very clean way out of this, but
3075 what I have chosen to do makes the common cases work: PCRE now takes note
3076 of whether there can be an explicit match for \r or \n anywhere in the
3077 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3078 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3079 pattern has explicit CR or LF references.
3081 29. Added (*CR) etc for changing newline setting at start of pattern.
3084 Version 7.2 19-Jun-07
3085 ---------------------
3087 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3088 which is apparently normally available under Windows.
3090 2. Re-jig the pcregrep tests with different newline settings in an attempt
3091 to make them independent of the local environment's newline setting.
3093 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3095 4. Some of the "internals" tests were previously cut out when the link size
3096 was not 2, because the output contained actual offsets. The recent new
3097 "Z" feature of pcretest means that these can be cut out, making the tests
3098 usable with all link sizes.
3100 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3101 stack recursion. This gives a massive performance boost under BSD, but just
3102 a small improvement under Linux. However, it saves one field in the frame
3103 in all cases.
3105 6. Added more features from the forthcoming Perl 5.10:
3107 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3108 recursion call. It refers to the nth most recently opened parentheses.
3110 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3111 to be opened parentheses.
3113 (c) Conditions that refer to capturing parentheses can be specified
3114 relatively, for example, (?(-2)... or (?(+3)...
3116 (d) \K resets the start of the current match so that everything before
3117 is not part of it.
3119 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3121 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3122 reference syntax.
3124 (g) (?| introduces a group in which the numbering of parentheses in each
3125 alternative starts with the same number.
3127 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3129 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3132 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3133 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3134 for detecting groups that can match an empty string.
3136 9. A pattern with a very large number of alternatives (more than several
3137 hundred) was running out of internal workspace during the pre-compile
3138 phase, where pcre_compile() figures out how much memory will be needed. A
3139 bit of new cunning has reduced the workspace needed for groups with
3140 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3141 workspace instead of running out of the 4096 that are available.
3143 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3145 11. Applied patch from Google to remove an optimization that didn't quite work.
3146 The report of the bug said:
3148 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3149 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3150 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3152 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3153 it matched the wrong number of bytes.
3156 Version 7.1 24-Apr-07
3157 ---------------------
3159 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3160 that is more "standard", making use of automake and other Autotools. There
3161 is some re-arrangement of the files and adjustment of comments consequent
3162 on this.
3164 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3165 for recursive directory scanning broke on some systems because the files
3166 are not scanned in any specific order and on different systems the order
3167 was different. A call to "sort" has been inserted into RunGrepTest for the
3168 approprate test as a short-term fix. In the longer term there may be an
3169 alternative.
3171 3. I had an email from Eric Raymond about problems translating some of PCRE's
3172 man pages to HTML (despite the fact that I distribute HTML pages, some
3173 people do their own conversions for various reasons). The problems
3174 concerned the use of low-level troff macros .br and .in. I have therefore
3175 removed all such uses from the man pages (some were redundant, some could
3176 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3177 HTML has been updated to handle .nf/.fi and to complain if it encounters
3178 .br or .in.
3180 4. Updated comments in configure.ac that get placed in config.h.in and also
3181 arranged for config.h to be included in the distribution, with the name
3182 config.h.generic, for the benefit of those who have to compile without
3183 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3185 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3186 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3187 makevp.bat for the current PCRE, using the additional files
3188 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3190 6. A Windows user reported a minor discrepancy with test 2, which turned out
3191 to be caused by a trailing space on an input line that had got lost in his
3192 copy. The trailing space was an accident, so I've just removed it.
3194 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3195 that is needed.
3197 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3198 as "const" (a) because they are and (b) because it helps the PHP
3199 maintainers who have recently made a script to detect big data structures
3200 in the php code that should be moved to the .rodata section. I remembered
3201 to update Builducptable as well, so it won't revert if ucptable.h is ever
3202 re-created.
3204 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3205 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3206 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3207 support is not required. This saves 1.5-2K of code, which is important in
3208 some applications.
3210 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3211 so as not to refer to the tables, even though these functions will never be
3212 called when UTF-8 support is disabled. Otherwise there are problems with a
3213 shared library.
3215 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3217 (a) It was defining its arguments as char * instead of void *.
3219 (b) It was assuming that all moves were upwards in memory; this was true
3220 a long time ago when I wrote it, but is no longer the case.
3222 The emulated memove() is provided for those environments that have neither
3223 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3224 is clearly not the case, as these two bugs were recently reported.
3226 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3227 and Detrail to create the HTML documentation, the .txt form of the man
3228 pages, and it removes trailing spaces from listed files. It also creates
3229 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3230 case, it wraps all the #defines with #ifndefs. This script should be run
3231 before "make dist".
3233 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3234 with Unicode property support.
3236 (a) For a maximizing quantifier, if the two different cases of the
3237 character were of different lengths in their UTF-8 codings (there are
3238 some cases like this - I found 11), and the matching function had to
3239 back up over a mixture of the two cases, it incorrectly assumed they
3240 were both the same length.
3242 (b) When PCRE was configured to use the heap rather than the stack for
3243 recursion during matching, it was not correctly preserving the data for
3244 the other case of a UTF-8 character when checking ahead for a match
3245 while processing a minimizing repeat. If the check also involved
3246 matching a wide character, but failed, corruption could cause an
3247 erroneous result when trying to check for a repeat of the original
3248 character.
3250 13. Some tidying changes to the testing mechanism:
3252 (a) The RunTest script now detects the internal link size and whether there
3253 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3254 values substituted by "configure". (The RunGrepTest script already did
3255 this for UTF-8.) The configure.ac script no longer substitutes the
3256 relevant variables.
3258 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3259 with length and offset values. This means that the output is different
3260 for different internal link sizes. Test 2 is skipped for link sizes
3261 other than 2 because of this, bypassing the problem. Unfortunately,
3262 there was also a test in test 3 (the locale tests) that used /B and
3263 failed for link sizes other than 2. Rather than cut the whole test out,
3264 I have added a new /Z option to pcretest that replaces the length and
3265 offset values with spaces. This is now used to make test 3 independent
3266 of link size. (Test 2 will be tidied up later.)
3268 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3269 segmentation fault instead of returning the appropriate error message.
3271 15. In multiline mode when the newline sequence was set to "any", the pattern
3272 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3273 This doesn't seem right; it now treats the CRLF combination as the line
3274 ending, and so does not match in that case. It's only a pattern such as ^$
3275 that would hit this one: something like ^ABC$ would have failed after \r
3276 and then tried again after \r\n.
3278 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3279 in an attempt to make files that differ only in their line terminators
3280 compare equal. This works on Linux.
3282 17. Under certain error circumstances pcregrep might try to free random memory
3283 as it exited. This is now fixed, thanks to valgrind.
3285 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3286 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3287 was because its rules for how to advance for /g after matching an empty
3288 string at the end of a line did not allow for this case. They now check for
3289 it specially.
3291 20. pcretest is supposed to handle patterns and data of any length, by
3292 extending its buffers when necessary. It was getting this wrong when the
3293 buffer for a data line had to be extended.
3295 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3296 CRLF as a newline sequence.
3298 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3299 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3300 I have nevertheless tidied it up.
3302 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3304 24. Added a man page for pcre-config.
3307 Version 7.0 19-Dec-06
3308 ---------------------
3310 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3311 moving to gcc 4.1.1.
3313 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3314 sys/time.h, which is documented as needed for this function. It doesn't
3315 seem to matter on Linux, but it showed up on some releases of OS X.
3317 3. It seems that there are systems where bytes whose values are greater than
3318 127 match isprint() in the "C" locale. The "C" locale should be the
3319 default when a C program starts up. In most systems, only ASCII printing
3320 characters match isprint(). This difference caused the output from pcretest
3321 to vary, making some of the tests fail. I have changed pcretest so that:
3323 (a) When it is outputting text in the compiled version of a pattern, bytes
3324 other than 32-126 are always shown as hex escapes.
3326 (b) When it is outputting text that is a matched part of a subject string,
3327 it does the same, unless a different locale has been set for the match
3328 (using the /L modifier). In this case, it uses isprint() to decide.
3330 4. Fixed a major bug that caused incorrect computation of the amount of memory
3331 required for a compiled pattern when options that changed within the
3332 pattern affected the logic of the preliminary scan that determines the
3333 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3334 that the computed length was too small. The symptoms of this bug were
3335 either the PCRE error "internal error: code overflow" from pcre_compile(),
3336 or a glibc crash with a message such as "pcretest: free(): invalid next
3337 size (fast)". Examples of patterns that provoked this bug (shown in
3338 pcretest format) are:
3340 /(?-x: )/x
3341 /(?x)(?-x: \s*#\s*)/
3342 /((?i)[\x{c0}])/8
3343 /(?i:[\x{c0}])/8
3345 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3346 is now done differently.
3348 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3349 wrapper classes; (b) implement a new function in the C++ scanner that is
3350 more efficient than the old way of doing things because it avoids levels of
3351 recursion in the regex matching; (c) add a paragraph to the documentation
3352 for the FullMatch() function.
3354 6. The escape sequence \n was being treated as whatever was defined as
3355 "newline". Not only was this contrary to the documentation, which states
3356 that \n is character 10 (hex 0A), but it also went horribly wrong when
3357 "newline" was defined as CRLF. This has been fixed.
3359 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3360 was being set to -1 for the "end of line" case (supposedly a value that no
3361 character can have). Though this value is never used (the check for end of
3362 line is "zero bytes in current character"), it caused compiler complaints.
3363 I've changed it to 0xffffffff.
3365 8. In pcre_version.c, the version string was being built by a sequence of
3366 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3367 string (as it is for production releases) called a macro with an empty
3368 argument. The C standard says the result of this is undefined. The gcc
3369 compiler treats it as an empty string (which was what was wanted) but it is
3370 reported that Visual C gives an error. The source has been hacked around to
3371 avoid this problem.
3373 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3374 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3375 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3376 of them did).
3378 10. Originally, pcretest opened its input and output without "b"; then I was
3379 told that "b" was needed in some environments, so it was added for release
3380 5.0 to both the input and output. (It makes no difference on Unix-like
3381 systems.) Later I was told that it is wrong for the input on Windows. I've
3382 now abstracted the modes into two macros, to make it easier to fiddle with
3383 them, and removed "b" from the input mode under Windows.
3385 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3387 12. Added -help and --help to pcretest as an official way of being reminded
3388 of the options.
3390 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3391 and pcrecpp.cc because they annoy compilers at high warning levels.
3393 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3395 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3396 scripts are not C programs :-) and which was not noticed because it works
3397 on Linux.
3399 16. pcretest is supposed to handle any length of pattern and data line (as one
3400 line or as a continued sequence of lines) by extending its input buffer if
3401 necessary. This feature was broken for very long pattern lines, leading to
3402 a string of junk being passed to pcre_compile() if the pattern was longer
3403 than about 50K.
3405 17. I have done a major re-factoring of the way pcre_compile() computes the
3406 amount of memory needed for a compiled pattern. Previously, there was code
3407 that made a preliminary scan of the pattern in order to do this. That was
3408 OK when PCRE was new, but as the facilities have expanded, it has become
3409 harder and harder to keep it in step with the real compile phase, and there
3410 have been a number of bugs (see for example, 4 above). I have now found a
3411 cunning way of running the real compile function in a "fake" mode that
3412 enables it to compute how much memory it would need, while actually only
3413 ever using a few hundred bytes of working memory and without too many
3414 tests of the mode. This should make future maintenance and development
3415 easier. A side effect of this work is that the limit of 200 on the nesting
3416 depth of parentheses has been removed (though this was never a serious
3417 limitation, I suspect). However, there is a downside: pcre_compile() now
3418 runs more slowly than before (30% or more, depending on the pattern). I
3419 hope this isn't a big issue. There is no effect on runtime performance.
3421 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3422 newline (only possible for the last line of a file) and it was a
3423 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3425 19. Added additional timing features to pcretest. (1) The -tm option now times
3426 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3427 separate command line item, by a number that specifies the number of
3428 repeats to use when timing. The default is 50000; this gives better
3429 precision, but takes uncomfortably long for very large patterns.
3431 20. Extended pcre_study() to be more clever in cases where a branch of a
3432 subpattern has no definite first character. For example, (a*|b*)[cd] would
3433 previously give no result from pcre_study(). Now it recognizes that the
3434 first character must be a, b, c, or d.
3436 21. There was an incorrect error "recursive call could loop indefinitely" if
3437 a subpattern (or the entire pattern) that was being tested for matching an
3438 empty string contained only one non-empty item after a nested subpattern.
3439 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3440 incorrectly, because the \d was being skipped in the check.
3442 22. The pcretest program now has a new pattern option /B and a command line
3443 option -b, which is equivalent to adding /B to every pattern. This causes
3444 it to show the compiled bytecode, without the additional information that
3445 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3446 is the same as /B/I).
3448 23. A new optimization is now able automatically to treat some sequences such
3449 as a*b as a*+b. More specifically, if something simple (such as a character
3450 or a simple class like \d) has an unlimited quantifier, and is followed by
3451 something that cannot possibly match the quantified thing, the quantifier
3452 is automatically "possessified".
3454 24. A recursive reference to a subpattern whose number was greater than 39
3455 went wrong under certain circumstances in UTF-8 mode. This bug could also
3456 have affected the operation of pcre_study().
3458 25. Realized that a little bit of performance could be had by replacing
3459 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3461 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3463 27. Possessive quantifiers such as a++ were previously implemented by turning
3464 them into atomic groups such as ($>a+). Now they have their own opcodes,
3465 which improves performance. This includes the automatically created ones
3466 from 23 above.
3468 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3469 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3470 the first matched character to be a colon. This applied both to named and
3471 numbered groups.
3473 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3475 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3476 building PCRE on Windows easier, so I have included it in the distribution.
3478 31. There is now a check in pcretest against a ridiculously large number being
3479 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3480 loop, the loop is abandoned.
3482 32. Forward references to subpatterns in conditions such as (?(2)...) where
3483 subpattern 2 is defined later cause pcre_compile() to search forwards in
3484 the pattern for the relevant set of parentheses. This search went wrong
3485 when there were unescaped parentheses in a character class, parentheses
3486 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3488 33. "Subroutine" calls and backreferences were previously restricted to
3489 referencing subpatterns earlier in the regex. This restriction has now
3490 been removed.
3492 34. Added a number of extra features that are going to be in Perl 5.10. On the
3493 whole, these are just syntactic alternatives for features that PCRE had
3494 previously implemented using the Python syntax or my own invention. The
3495 other formats are all retained for compatibility.
3497 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3498 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3499 also .NET compatible.
3501 (b) A recursion or subroutine call to a named group can now be defined as
3502 (?&name) as well as (?P>name).
3504 (c) A backreference to a named group can now be defined as \k<name> or
3505 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3506 5.10, are also .NET compatible.
3508 (d) A conditional reference to a named group can now use the syntax
3509 (?(<name>) or (?('name') as well as (?(name).
3511 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3512 groups (named and numbered) that are never evaluated inline, but can be
3513 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3514 is always false. There may be only one alternative in such a group.
3516 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3517 as the simple (?(R). The condition is true only if the most recent
3518 recursion is that of the given number or name. It does not search out
3519 through the entire recursion stack.
3521 (g) The escape \gN or \g{N} has been added, where N is a positive or
3522 negative number, specifying an absolute or relative reference.
3524 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3525 some "unreachable code" warnings.
3527 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3528 things, this adds five new scripts.
3530 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3531 There were also incompatibilities regarding the handling of \Q..\E inside
3532 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3533 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3535 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3536 matches an empty string, and forcibly breaks the loop. There were bugs in
3537 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3538 against aaaa the result was just "a" rather than "aaaa", for example. Two
3539 separate and independent bugs (that affected different cases) have been
3540 fixed.
3542 39. Refactored the code to abolish the use of different opcodes for small
3543 capturing bracket numbers. This is a tidy that I avoided doing when I
3544 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3545 The new approach is not only tidier, it makes it possible to reduce the
3546 memory needed to fix the previous bug (38).
3548 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3549 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3550 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3551 mode.
3553 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3554 report.
3556 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3557 copy construction and assignment in the C++ wrapper.
3559 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3560 couple of bugs that could have given wrong results in the "--newline=crlf"
3561 case.
3563 44. Added a number of casts and did some reorganization of signed/unsigned int
3564 variables following suggestions from Dair Grant. Also renamed the variable
3565 "this" as "item" because it is a C++ keyword.
3567 45. Arranged for dftables to add
3569 #include "pcre_internal.h"
3571 to pcre_chartables.c because without it, gcc 4.x may remove the array
3572 definition from the final binary if PCRE is built into a static library and
3573 dead code stripping is activated.
3575 46. For an unanchored pattern, if a match attempt fails at the start of a
3576 newline sequence, and the newline setting is CRLF or ANY, and the next two
3577 characters are CRLF, advance by two characters instead of one.
3580 Version 6.7 04-Jul-06
3581 ---------------------
3583 1. In order to handle tests when input lines are enormously long, pcretest has
3584 been re-factored so that it automatically extends its buffers when
3585 necessary. The code is crude, but this _is_ just a test program. The
3586 default size has been increased from 32K to 50K.
3588 2. The code in pcre_study() was using the value of the re argument before
3589 testing it for NULL. (Of course, in any sensible call of the function, it
3590 won't be NULL.)
3592 3. The memmove() emulation function in pcre_internal.h, which is used on
3593 systems that lack both memmove() and bcopy() - that is, hardly ever -
3594 was missing a "static" storage class specifier.
3596 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3597 containing an extended class (one that cannot be represented by a bitmap
3598 because it contains high-valued characters or Unicode property items, e.g.
3599 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3600 pattern, but PCRE should not loop if you do not (it no longer does).
3601 [Detail: two cases were found: (a) a repeated subpattern containing an
3602 extended class; (b) a recursive reference to a subpattern that followed a
3603 previous extended class. It wasn't skipping over the extended class
3604 correctly when UTF-8 mode was not set.]
3606 5. A negated single-character class was not being recognized as fixed-length
3607 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3608 compile error "lookbehind assertion is not fixed length".
3610 6. The RunPerlTest auxiliary script was showing an unexpected difference
3611 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3612 write a Perl script that can interpret lines of an input file either as
3613 byte characters or as UTF-8, which is what "perltest" was being required to
3614 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3615 can't do is switch easily at run time between having the "use utf8;" pragma
3616 or not. In the end, I fudged it by using the RunPerlTest script to insert
3617 "use utf8;" explicitly for the UTF-8 tests.
3619 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3620 the end of the subject string, contrary to the documentation and to what
3621 Perl does. This was true of both matching functions. Now it matches only at
3622 the start of the subject and immediately after *internal* newlines.
3624 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3625 a pointer to an int instead of a pointer to an unsigned long int. This
3626 caused problems on 64-bit systems.
3628 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3629 instance of the 'standard' template library not being so standard".
3631 10. There was no check on the number of named subpatterns nor the maximum
3632 length of a subpattern name. The product of these values is used to compute
3633 the size of the memory block for a compiled pattern. By supplying a very
3634 long subpattern name and a large number of named subpatterns, the size
3635 computation could be caused to overflow. This is now prevented by limiting
3636 the length of names to 32 characters, and the number of named subpatterns
3637 to 10,000.
3639 11. Subpatterns that are repeated with specific counts have to be replicated in
3640 the compiled pattern. The size of memory for this was computed from the
3641 length of the subpattern and the repeat count. The latter is limited to
3642 65535, but there was no limit on the former, meaning that integer overflow
3643 could in principle occur. The compiled length of a repeated subpattern is
3644 now limited to 30,000 bytes in order to prevent this.
3646 12. Added the optional facility to have named substrings with the same name.
3648 13. Added the ability to use a named substring as a condition, using the
3649 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3650 are numbers (not recommended). Forward references are permitted.
3652 14. Added forward references in named backreferences (if you see what I mean).
3654 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3655 pattern could run off the end of the subject. For example, the pattern
3656 "(?s)(.{1,5})"8 did this with the subject "ab".
3658 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3659 PCRE_CASELESS was set when matching characters that were quantified with ?
3660 or *.
3662 17. A character class other than a single negated character that had a minimum
3663 but no maximum quantifier - for example [ab]{6,} - was not handled
3664 correctly by pce_dfa_exec(). It would match only one character.
3666 18. A valid (though odd) pattern that looked like a POSIX character
3667 class but used an invalid character after [ (for example [[,abc,]]) caused
3668 pcre_compile() to give the error "Failed: internal error: code overflow" or
3669 in some cases to crash with a glibc free() error. This could even happen if
3670 the pattern terminated after [[ but there just happened to be a sequence of
3671 letters, a binary zero, and a closing ] in the memory that followed.
3673 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3674 over the years. Originally (before any Unicode support), just the bottom 8
3675 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3676 output from "man perlunicode" includes this:
3678 The regular expression compiler produces polymorphic opcodes. That
3679 is, the pattern adapts to the data and automatically switches to
3680 the Unicode character scheme when presented with Unicode data--or
3681 instead uses a traditional byte scheme when presented with byte
3682 data.
3684 Sadly, a wide octal escape does not cause a switch, and in a string with
3685 no other multibyte characters, these octal escapes are treated as before.
3686 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3687 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3688 Unicode string.
3690 I have not perpetrated such confusion in PCRE. Up till now, it took just
3691 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3692 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3693 translate to the appropriate multibyte character.
3695 29. Applied some refactoring to reduce the number of warnings from Microsoft
3696 and Borland compilers. This has included removing the fudge introduced
3697 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3698 a warning about an unused variable.
3700 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3701 characters since release 4.0, because Perl (from release 5.004) does not.
3702 [Or at least, is documented not to: some releases seem to be in conflict
3703 with the documentation.] However, when a pattern was studied with
3704 pcre_study() and all its branches started with \s, PCRE still included VT
3705 as a possible starting character. Of course, this did no harm; it just
3706 caused an unnecessary match attempt.
3708 22. Removed a now-redundant internal flag bit that recorded the fact that case
3709 dependency changed within the pattern. This was once needed for "required
3710 byte" processing, but is no longer used. This recovers a now-scarce options
3711 bit. Also moved the least significant internal flag bit to the most-
3712 significant bit of the word, which was not previously used (hangover from
3713 the days when it was an int rather than a uint) to free up another bit for
3714 the future.
3716 23. Added support for CRLF line endings as well as CR and LF. As well as the
3717 default being selectable at build time, it can now be changed at runtime
3718 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3719 specify that it is scanning data with non-default line endings.
3721 24. Changed the definition of CXXLINK to make it agree with the definition of
3722 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3724 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3725 recursions. This makes a big different to stack usage for some patterns.
3727 26. If a subpattern containing a named recursion or subroutine reference such
3728 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3729 the space required for the compiled pattern went wrong and gave too small a
3730 value. Depending on the environment, this could lead to "Failed: internal
3731 error: code overflow at offset 49" or "glibc detected double free or
3732 corruption" errors.
3734 27. Applied patches from Google (a) to support the new newline modes and (b) to
3735 advance over multibyte UTF-8 characters in GlobalReplace.
3737 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3738 difference for some implementation of PCRE in some Windows version.
3740 29. Added some extra testing facilities to pcretest:
3742 \q<number> in a data line sets the "match limit" value
3743 \Q<number> in a data line sets the "match recursion limt" value
3744 -S <number> sets the stack size, where <number> is in megabytes
3746 The -S option isn't available for Windows.
3749 Version 6.6 06-Feb-06
3750 ---------------------
3752 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3753 in pcreposix.h. I have copied the definition from pcre.h.
3755 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3756 because pcre.h is no longer a built file.
3758 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3759 not normally included in the compiled code.
3762 Version 6.5 01-Feb-06
3763 ---------------------
3765 1. When using the partial match feature with pcre_dfa_exec(), it was not
3766 anchoring the second and subsequent partial matches at the new starting
3767 point. This could lead to incorrect results. For example, with the pattern
3768 /1234/, partially matching against "123" and then "a4" gave a match.
3770 2. Changes to pcregrep:
3772 (a) All non-match returns from pcre_exec() were being treated as failures
3773 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3774 error message is output. Some extra information is given for the
3776 probably the only errors that are likely to be caused by users (by
3777 specifying a regex that has nested indefinite repeats, for instance).
3778 If there are more than 20 of these errors, pcregrep is abandoned.
3780 (b) A binary zero was treated as data while matching, but terminated the
3781 output line if it was written out. This has been fixed: binary zeroes
3782 are now no different to any other data bytes.
3784 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3785 used to set a locale for matching. The --locale=xxxx long option has
3786 been added (no short equivalent) to specify a locale explicitly on the
3787 pcregrep command, overriding the environment variables.
3789 (d) When -B was used with -n, some line numbers in the output were one less
3790 than they should have been.
3792 (e) Added the -o (--only-matching) option.
3794 (f) If -A or -C was used with -c (count only), some lines of context were
3795 accidentally printed for the final match.
3797 (g) Added the -H (--with-filename) option.
3799 (h) The combination of options -rh failed to suppress file names for files
3800 that were found from directory arguments.
3802 (i) Added the -D (--devices) and -d (--directories) options.
3804 (j) Added the -F (--fixed-strings) option.
3806 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3808 (l) Added the --colo(u)r option.
3810 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3811 is not present by default.
3813 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3814 items such as (?R) or (?1), when the recursion could match a number of
3815 alternatives. If it matched one of the alternatives, but subsequently,
3816 outside the recursion, there was a failure, the code tried to back up into
3817 the recursion. However, because of the way PCRE is implemented, this is not
3818 possible, and the result was an incorrect result from the match.
3820 In order to prevent this happening, the specification of recursion has
3821 been changed so that all such subpatterns are automatically treated as
3822 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3824 4. I had overlooked the fact that, in some locales, there are characters for
3825 which isalpha() is true but neither isupper() nor islower() are true. In
3826 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3827 and ordfeminine) are like this. This affected the treatment of \w and \W
3828 when they appeared in character classes, but not when they appeared outside
3829 a character class. The bit map for "word" characters is now created
3830 separately from the results of isalnum() instead of just taking it from the
3831 upper, lower, and digit maps. (Plus the underscore character, of course.)
3833 5. The above bug also affected the handling of POSIX character classes such as
3834 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3835 permanent tables. Instead, the bit maps for such a class were previously
3836 created as the appropriate unions of the upper, lower, and digit bitmaps.
3837 Now they are created by subtraction from the [[:word:]] class, which has
3838 its own bitmap.
3840 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3841 It is created by subtracting the vertical space characters (\x09, \x0a,
3842 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3843 subtraction was done in the overall bitmap for a character class, meaning
3844 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3845 be recognized. This bug has been fixed.
3847 7. Patches from the folks at Google:
3849 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3850 real life, but is still worth protecting against".
3852 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3853 regular expressions".
3855 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3856 have it.
3858 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3859 "configure" and the latter not, in order to fix a problem somebody had
3860 with compiling the Arg class on HP-UX.
3862 (e) Improve the error-handling of the C++ wrapper a little bit.
3864 (f) New tests for checking recursion limiting.
3866 8. The pcre_memmove() function, which is used only if the environment does not
3867 have a standard memmove() function (and is therefore rarely compiled),
3868 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3869 returning a result (though PCRE never actually uses the result).
3871 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3872 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3873 returned instead of calling malloc() with an overflowing number that would
3874 most likely cause subsequent chaos.
3876 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3878 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3879 with this option is matched, the nmatch and pmatch options of regexec() are
3880 ignored.
3882 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3883 provided in case anyone wants to the the POSIX interface with UTF-8
3884 strings.
3886 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3887 C++ linking (needed for some HP-UX environments).
3889 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3890 (unused parameter) and in the pcre_printint() function (omitted "default"
3891 switch label when the default is to do nothing).
3893 15. Added some code to make it possible, when PCRE is compiled as a C++
3894 library, to replace subject pointers for pcre_exec() with a smart pointer
3895 class, thus making it possible to process discontinuous strings.
3897 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3898 much the same function. They were added by different people who were trying
3899 to make PCRE easy to compile on non-Unix systems. It has been suggested
3900 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3901 for compiling on Windows systems. I have therefore replaced it with
3902 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3903 defaults to "extern" for C or "extern C" for C++, which works fine on
3904 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3905 SCOPE with something explicit in config.h. In addition:
3907 (a) pcreposix.h still had just "extern" instead of either of these macros;
3908 I have replaced it with PCRE_DATA_SCOPE.
3910 (b) Functions such as _pcre_xclass(), which are internal to the library,
3911 but external in the C sense, all had PCRE_EXPORT in their definitions.
3912 This is apparently wrong for the Windows case, so I have removed it.
3913 (It makes no difference on Unix-like systems.)
3915 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3916 of recursive calls to match(). This is different to MATCH_LIMIT because
3917 that limits the total number of calls to match(), not all of which increase
3918 the depth of recursion. Limiting the recursion depth limits the amount of
3919 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3920 when PCRE is compiled, and changed at run time. A patch from Google adds
3921 this functionality to the C++ interface.
3923 18. Changes to the handling of Unicode character properties:
3925 (a) Updated the table to Unicode 4.1.0.
3927 (b) Recognize characters that are not in the table as "Cn" (undefined).
3929 (c) I revised the way the table is implemented to a much improved format
3930 which includes recognition of ranges. It now supports the ranges that
3931 are defined in UnicodeData.txt, and it also amalgamates other
3932 characters into ranges. This has reduced the number of entries in the
3933 table from around 16,000 to around 3,000, thus reducing its size
3934 considerably. I realized I did not need to use a tree structure after
3935 all - a binary chop search is just as efficient. Having reduced the
3936 number of entries, I extended their size from 6 bytes to 8 bytes to
3937 allow for more data.
3939 (d) Added support for Unicode script names via properties such as \p{Han}.
3941 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3942 matching that character.
3944 20. When matching a repeated Unicode property with a minimum greater than zero,
3945 (for example \pL{2,}), PCRE could look past the end of the subject if it
3946 reached it while seeking the minimum number of characters. This could
3947 happen only if some of the characters were more than one byte long, because
3948 there is a check for at least the minimum number of bytes.
3950 21. Refactored the implementation of \p and \P so as to be more general, to
3951 allow for more different types of property in future. This has changed the
3952 compiled form incompatibly. Anybody with saved compiled patterns that use
3953 \p or \P will have to recompile them.
3955 22. Added "Any" and "L&" to the supported property types.
3957 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3958 but give a compile time error if the value is greater than 0xff.
3960 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3961 accidentally not being installed or uninstalled.
3963 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3964 made were to insert the current release number. This seemed silly, because
3965 it made things harder for people building PCRE on systems that don't run
3966 "configure". I have turned pcre.h into a distributed file, no longer built
3967 by "configure", with the version identification directly included. There is
3968 no longer a pcre.h.in file.
3970 However, this change necessitated a change to the pcre-config script as
3971 well. It is built from pcre-config.in, and one of the substitutions was the
3972 release number. I have updated configure.ac so that ./configure now finds
3973 the release number by grepping pcre.h.
3975 26. Added the ability to run the tests under valgrind.
3978 Version 6.4 05-Sep-05
3979 ---------------------
3981 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3982 "--" to be printed when multiple files were scanned, even when none of the
3983 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3984 consider it to be a bug, and have restored the previous behaviour.
3986 2. A couple of code tidies to get rid of compiler warnings.
3988 3. The pcretest program used to cheat by referring to symbols in the library
3989 whose names begin with _pcre_. These are internal symbols that are not
3990 really supposed to be visible externally, and in some environments it is
3991 possible to suppress them. The cheating is now confined to including
3992 certain files from the library's source, which is a bit cleaner.
3994 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3995 file's purpose clearer.
3997 5. Reorganized pcre_ucp_findchar().
4000 Version 6.3 15-Aug-05
4001 ---------------------
4003 1. The file libpcre.pc.in did not have general read permission in the tarball.
4005 2. There were some problems when building without C++ support:
4007 (a) If C++ support was not built, "make install" and "make test" still
4008 tried to test it.
4010 (b) There were problems when the value of CXX was explicitly set. Some
4011 changes have been made to try to fix these, and ...
4013 (c) --disable-cpp can now be used to explicitly disable C++ support.
4015 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4016 backslash in a target when C++ was disabled. This confuses some
4017 versions of "make", apparently. Using an intermediate variable solves
4018 this. (Same for CPP_LOBJ.)
4020 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4021 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4022 necessary on certain architectures.
4024 4. Added a setting of -export-symbols-regex to the link command to remove
4025 those symbols that are exported in the C sense, but actually are local
4026 within the library, and not documented. Their names all begin with
4027 "_pcre_". This is not a perfect job, because (a) we have to except some
4028 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4029 available (and never for static libraries). I have made a note to try to
4030 find a way round (a) in the future.
4033 Version 6.2 01-Aug-05
4034 ---------------------
4036 1. There was no test for integer overflow of quantifier values. A construction
4037 such as {1111111111111111} would give undefined results. What is worse, if
4038 a minimum quantifier for a parenthesized subpattern overflowed and became
4039 negative, the calculation of the memory size went wrong. This could have
4040 led to memory overwriting.
4042 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4044 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4045 operating environments where this matters.
4047 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4048 PCRE options from within the C++ wrapper.
4050 5. Named capturing subpatterns were not being correctly counted when a pattern
4051 was compiled. This caused two problems: (a) If there were more than 100
4052 such subpatterns, the calculation of the memory needed for the whole
4053 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4054 back references of the form \12, where the number was greater than 9, were
4055 not recognized as back references, even though there were sufficient
4056 previous subpatterns.
4058 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4059 versions of gcc, e.g. 2.95.4.
4062 Version 6.1 21-Jun-05
4063 ---------------------
4065 1. There was one reference to the variable "posix" in pcretest.c that was not
4066 surrounded by "#if !defined NOPOSIX".
4068 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4069 the cross-check on the old pcre_info() function, for the benefit of the
4070 cut-down version of PCRE that is currently imported into Exim.
4072 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4073 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4074 patterns that start (?i) but I don't think that matters. The use of (?i) is
4075 just an example; this all applies to the other options as well.
4077 4. Since libtool seems to echo the compile commands it is issuing, the output
4078 from "make" can be reduced a bit by putting "@" in front of each libtool
4079 compile command.
4081 5. Patch from the folks at Google for configure.in to be a bit more thorough
4082 in checking for a suitable C++ installation before trying to compile the
4083 C++ stuff. This should fix a reported problem when a compiler was present,
4084 but no suitable headers.
4086 6. The man pages all had just "PCRE" as their title. I have changed them to
4087 be the relevant file name. I have also arranged that these names are
4088 retained in the file doc/pcre.txt, which is a concatenation in text format
4089 of all the man pages except the little individual ones for each function.
4091 7. The NON-UNIX-USE file had not been updated for the different set of source
4092 files that come with release 6. I also added a few comments about the C++
4093 wrapper.
4096 Version 6.0 07-Jun-05
4097 ---------------------
4099 1. Some minor internal re-organization to help with my DFA experiments.
4101 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4102 didn't matter for the library itself when fully configured, but did matter
4103 when compiling without UCP support, or within Exim, where the ucp files are
4104 not imported.
4106 3. Refactoring of the library code to split up the various functions into
4107 different source modules. The addition of the new DFA matching code (see
4108 below) to a single monolithic source would have made it really too
4109 unwieldy, quite apart from causing all the code to be include in a
4110 statically linked application, when only some functions are used. This is
4111 relevant even without the DFA addition now that patterns can be compiled in
4112 one application and matched in another.
4114 The downside of splitting up is that there have to be some external
4115 functions and data tables that are used internally in different modules of
4116 the library but which are not part of the API. These have all had their
4117 names changed to start with "_pcre_" so that they are unlikely to clash
4118 with other external names.
4120 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4121 a different (DFA) algorithm. Although it is slower than the original
4122 function, it does have some advantages for certain types of matching
4123 problem.
4125 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4126 including restarting after a partial match.
4128 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4129 defined when compiling for Windows was sent to me. I have put it into the
4130 code, though I have no means of testing or verifying it.
4132 7. Added the pcre_refcount() auxiliary function.
4134 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4135 match before or at the first newline in the subject string. In pcretest,
4136 the /f option on a pattern can be used to set this.
4138 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4139 would behave wrongly. This has been present in PCRE since release 4.0.
4141 10. A number of changes to the pcregrep command:
4143 (a) Refactored how -x works; insert ^(...)$ instead of setting
4144 PCRE_ANCHORED and checking the length, in preparation for adding
4145 something similar for -w.
4147 (b) Added the -w (match as a word) option.
4149 (c) Refactored the way lines are read and buffered so as to have more
4150 than one at a time available.
4152 (d) Implemented a pcregrep test script.
4154 (e) Added the -M (multiline match) option. This allows patterns to match
4155 over several lines of the subject. The buffering ensures that at least
4156 8K, or the rest of the document (whichever is the shorter) is available
4157 for matching (and similarly the previous 8K for lookbehind assertions).
4159 (f) Changed the --help output so that it now says
4161 -w, --word-regex(p)
4163 instead of two lines, one with "regex" and the other with "regexp"
4164 because that confused at least one person since the short forms are the
4165 same. (This required a bit of code, as the output is generated
4166 automatically from a table. It wasn't just a text change.)
4168 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4169 option but starts with a hyphen. Could be a pattern or a path name
4170 starting with a hyphen, for instance.
4172 (h) "-" can be given as a file name to represent stdin.
4174 (i) When file names are being printed, "(standard input)" is used for
4175 the standard input, for compatibility with GNU grep. Previously
4176 "<stdin>" was used.
4178 (j) The option --label=xxx can be used to supply a name to be used for
4179 stdin when file names are being printed. There is no short form.
4181 (k) Re-factored the options decoding logic because we are going to add
4182 two more options that take data. Such options can now be given in four
4183 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4185 (l) Added the -A, -B, and -C options for requesting that lines of context
4186 around matches be printed.
4188 (m) Added the -L option to print the names of files that do not contain
4189 any matching lines, that is, the complement of -l.
4191 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4192 continue to scan other files.
4194 (o) The -s option was incorrectly implemented. For compatibility with other
4195 greps, it now suppresses the error message for a non-existent or non-
4196 accessible file (but not the return code). There is a new option called
4197 -q that suppresses the output of matching lines, which was what -s was
4198 previously doing.
4200 (p) Added --include and --exclude options to specify files for inclusion
4201 and exclusion when recursing.
4203 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4204 Hopefully, it now does.
4206 12. Missing cast in pcre_study().
4208 13. Added an "uninstall" target to the makefile.
4210 14. Replaced "extern" in the function prototypes in Makefile.in with
4211 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4212 world, but is set differently for Windows.
4214 15. Added a second compiling function called pcre_compile2(). The only
4215 difference is that it has an extra argument, which is a pointer to an
4216 integer error code. When there is a compile-time failure, this is set
4217 non-zero, in addition to the error test pointer being set to point to an
4218 error message. The new argument may be NULL if no error number is required
4219 (but then you may as well call pcre_compile(), which is now just a
4220 wrapper). This facility is provided because some applications need a
4221 numeric error indication, but it has also enabled me to tidy up the way
4222 compile-time errors are handled in the POSIX wrapper.
4224 16. Added VPATH=.libs to the makefile; this should help when building with one
4225 prefix path and installing with another. (Or so I'm told by someone who
4226 knows more about this stuff than I do.)
4228 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4229 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4230 match everything, including newlines. This is not POSIX-compatible, but
4231 somebody wanted the feature. From pcretest it can be activated by using
4232 both the P and the s flags.
4234 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4236 19. libpcre.pc was being incorrectly installed as executable.
4238 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4239 it now also looks for '\r' so that it will work unmodified on Windows.
4241 21. Added Google's contributed C++ wrapper to the distribution.
4243 22. Added some untidy missing memory free() calls in pcretest, to keep
4244 Electric Fence happy when testing.
4248 Version 5.0 13-Sep-04
4249 ---------------------
4251 1. Internal change: literal characters are no longer packed up into items
4252 containing multiple characters in a single byte-string. Each character
4253 is now matched using a separate opcode. However, there may be more than one
4254 byte in the character in UTF-8 mode.
4256 2. The pcre_callout_block structure has two new fields: pattern_position and
4257 next_item_length. These contain the offset in the pattern to the next match
4258 item, and its length, respectively.
4260 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4261 insertion of callouts before each pattern item. Added the /C option to
4262 pcretest to make use of this.
4264 4. On the advice of a Windows user, the lines
4266 #if defined(_WIN32) || defined(WIN32)
4267 _setmode( _fileno( stdout ), 0x8000 );
4268 #endif /* defined(_WIN32) || defined(WIN32) */
4270 have been added to the source of pcretest. This apparently does useful
4271 magic in relation to line terminators.
4273 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4274 for the benefit of those environments where the "b" makes a difference.
4276 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4277 to know about it. I have put a hack into configure.in that adds in code
4278 to set GCC=yes if CC=icc. This seems to end up at a point in the
4279 generated configure script that is early enough to affect the setting of
4280 compiler options, which is what is needed, but I have no means of testing
4281 whether it really works. (The user who reported this had patched the
4282 generated configure script, which of course I cannot do.)
4284 LATER: After change 22 below (new libtool files), the configure script
4285 seems to know about icc (and also ecc). Therefore, I have commented out
4286 this hack in configure.in.
4288 7. Added support for pkg-config (2 patches were sent in).
4290 8. Negated POSIX character classes that used a combination of internal tables
4291 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4292 [[:^ascii]]. Typically, they would match almost any characters. The other
4293 POSIX classes were not broken in this way.
4295 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4296 to find the match, as PCRE was deluded into thinking that the match had to
4297 start at the start point or following a newline. The same bug applied to
4298 patterns with negative forward assertions or any backward assertions
4299 preceding ".*" at the start, unless the pattern required a fixed first
4300 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4302 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4303 starting at the last subject character, bytes beyond the end of the subject
4304 string were read.
4306 11. Renamed the variable "class" as "classbits" to make life easier for C++
4307 users. (Previously there was a macro definition, but it apparently wasn't
4308 enough.)
4310 12. Added the new field "tables" to the extra data so that tables can be passed
4311 in at exec time, or the internal tables can be re-selected. This allows
4312 a compiled regex to be saved and re-used at a later time by a different
4313 program that might have everything at different addresses.
4315 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4316 -R library as well as a -L library.
4318 14. The debugging options of pcretest (-d on the command line or D on a
4319 pattern) showed incorrect output for anything following an extended class
4320 that contained multibyte characters and which was followed by a quantifier.
4322 15. Added optional support for general category Unicode character properties
4323 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4324 support. It adds about 90K to the size of the library. The meanings of the
4325 inbuilt class escapes such as \d and \s have NOT been changed.
4327 16. Updated pcredemo.c to include calls to free() to release the memory for the
4328 compiled pattern.
4330 17. The generated file chartables.c was being created in the source directory
4331 instead of in the building directory. This caused the build to fail if the
4332 source directory was different from the building directory, and was
4333 read-only.
4335 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4336 file. No doubt somebody will tell me if they don't make sense... Also added
4337 Dan Mooney's comments about building on OpenVMS.
4339 19. Added support for partial matching via the PCRE_PARTIAL option for
4340 pcre_exec() and the \P data escape in pcretest.
4342 20. Extended pcretest with 3 new pattern features:
4344 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4345 write the compiled pattern to the file whose name is "rest-of-line".
4346 This is a straight binary dump of the data, with the saved pointer to
4347 the character tables forced to be NULL. The study data, if any, is
4348 written too. After writing, pcretest reads a new pattern.
4350 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4351 compiled pattern from the given file. There must not be any
4352 occurrences of "<" in the file name (pretty unlikely); if there are,
4353 pcretest will instead treat the initial "<" as a pattern delimiter.
4354 After reading in the pattern, pcretest goes on to read data lines as
4355 usual.
4357 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4358 and 16-bit fields in a compiled pattern, to simulate a pattern that
4359 was compiled on a host of opposite endianness.
4361 21. The pcre-exec() function can now cope with patterns that were compiled on
4362 hosts of opposite endianness, with this restriction:
4364 As for any compiled expression that is saved and used later, the tables
4365 pointer field cannot be preserved; the extra_data field in the arguments
4366 to pcre_exec() should be used to pass in a tables address if a value
4367 other than the default internal tables were used at compile time.
4369 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4370 now diagnosed as an error. Previously, most of the time, a negative number
4371 would have been treated as zero, but if in addition "ovector" was passed as
4372 NULL, a crash could occur.
4374 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4375 new versions from the libtool 1.5 distribution (the last one is a copy of
4376 a file called libtool.m4). This seems to have fixed the need to patch
4377 "configure" to support Darwin 1.3 (which I used to do). However, I still
4378 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4379 workstation).
4381 24. Changed the PCRE licence to be the more standard "BSD" licence.
4384 Version 4.5 01-Dec-03
4385 ---------------------
4387 1. There has been some re-arrangement of the code for the match() function so
4388 that it can be compiled in a version that does not call itself recursively.
4389 Instead, it keeps those local variables that need separate instances for
4390 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4391 needs to "recurse". Keeping track of where control must go is done by means
4392 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4393 hide most of the details from the main code, and operates only if
4394 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4395 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4396 operating.
4398 To make it easier for callers to provide specially tailored get/free
4399 functions for this usage, two new functions, pcre_stack_malloc, and
4400 pcre_stack_free, are used. They are always called in strict stacking order,
4401 and the size of block requested is always the same.
4403 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4404 PCRE has been compiled to use the stack or the heap for recursion. The
4405 -C option of pcretest uses this to show which version is compiled.
4407 A new data escape \S, is added to pcretest; it causes the amounts of store
4408 obtained and freed by both kinds of malloc/free at match time to be added
4409 to the output.
4411 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4412 what's available on my current Linux desktop machine.
4414 3. When matching a UTF-8 string, the test for a valid string at the start has
4415 been extended. If start_offset is not zero, PCRE now checks that it points
4416 to a byte that is the start of a UTF-8 character. If not, it returns
4417 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4418 this is necessary because there may be backward assertions in the pattern.
4419 When matching the same subject several times, it may save resources to use
4420 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4422 4. The code for checking the validity of UTF-8 strings has been tightened so
4423 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4424 containing "overlong sequences".
4426 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4427 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4428 should have been "&", but it just so happened that all the cases this let
4429 through by mistake were picked up later in the function.
4431 6. I had used a variable called "isblank" - this is a C99 function, causing
4432 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4434 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4435 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4436 so the version is shown; (c) stop "make test" echoing "./RunTest".
4438 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4440 9. The prototype for memmove() for systems that don't have it was using
4441 size_t, but the inclusion of the header that defines size_t was later. I've
4442 moved the #includes for the C headers earlier to avoid this.
4444 10. Added some adjustments to the code to make it easier to compiler on certain
4445 special systems:
4447 (a) Some "const" qualifiers were missing.
4448 (b) Added the macro EXPORT before all exported functions; by default this
4449 is defined to be empty.
4450 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4451 that it reads its output file name as an argument instead of writing
4452 to the standard output and assuming this can be redirected.
4454 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4455 class containing characters with values greater than 255, PCRE compilation
4456 went into a loop.
4458 12. A recursive reference to a subpattern that was within another subpattern
4459 that had a minimum quantifier of zero caused PCRE to crash. For example,
4460 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4461 recursion. If the recursively-called subpattern itself had a zero repeat,
4462 that was OK.
4464 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4465 buffer into which it was copied (for escape processing) was still set at
4466 1024, so long lines caused crashes.
4468 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4469 "internal error: code overflow...". This applied to any character class
4470 that was followed by a possessive quantifier.
4472 15. Modified the Makefile to add libpcre.la as a prerequisite for
4473 libpcreposix.la because I was told this is needed for a parallel build to
4474 work.
4476 16. If a pattern that contained .* following optional items at the start was
4477 studied, the wrong optimizing data was generated, leading to matching
4478 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4479 matching string must start with a or b or c. The correct conclusion for
4480 this pattern is that a match can start with any character.
4483 Version 4.4 13-Aug-03
4484 ---------------------
4486 1. In UTF-8 mode, a character class containing characters with values between
4487 127 and 255 was not handled correctly if the compiled pattern was studied.
4488 In fixing this, I have also improved the studying algorithm for such
4489 classes (slightly).
4491 2. Three internal functions had redundant arguments passed to them. Removal
4492 might give a very teeny performance improvement.
4494 3. Documentation bug: the value of the capture_top field in a callout is *one
4495 more than* the number of the hightest numbered captured substring.
4497 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4498 in incorrectly linking with a previously installed version. They now link
4499 explicitly with libpcre.la.
4501 5. configure.in no longer needs to recognize Cygwin specially.
4503 6. A problem in pcre.in for Windows platforms is fixed.
4505 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4506 pcretest, it used to include the size of the study block as part of its
4507 output. Unfortunately, the structure contains a field that has a different
4508 size on different hardware architectures. This meant that the tests that
4509 showed this size failed. As the block is currently always of a fixed size,
4510 this information isn't actually particularly useful in pcretest output, so
4511 I have just removed it.
4513 8. Three pre-processor statements accidentally did not start in column 1.
4514 Sadly, there are *still* compilers around that complain, even though
4515 standard C has not required this for well over a decade. Sigh.
4517 9. In pcretest, the code for checking callouts passed small integers in the
4518 callout_data field, which is a void * field. However, some picky compilers
4519 complained about the casts involved for this on 64-bit systems. Now
4520 pcretest passes the address of the small integer instead, which should get
4521 rid of the warnings.
4523 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4524 both compile and run time, and gives an error if an invalid UTF-8 sequence
4525 is found. There is a option for disabling this check in cases where the
4526 string is known to be correct and/or the maximum performance is wanted.
4528 11. In response to a bug report, I changed one line in Makefile.in from
4530 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4531 to
4532 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4534 to look similar to other lines, but I have no way of telling whether this
4535 is the right thing to do, as I do not use Windows. No doubt I'll get told
4536 if it's wrong...
4539 Version 4.3 21-May-03
4540 ---------------------
4542 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4543 Makefile.
4545 2. Some refactoring to improve the quality of the code:
4547 (i) The utf8_table... variables are now declared "const".
4549 (ii) The code for \cx, which used the "case flipping" table to upper case
4550 lower case letters, now just substracts 32. This is ASCII-specific,
4551 but the whole concept of \cx is ASCII-specific, so it seems
4552 reasonable.
4554 (iii) PCRE was using its character types table to recognize decimal and
4555 hexadecimal digits in the pattern. This is silly, because it handles
4556 only 0-9, a-f, and A-F, but the character types table is locale-
4557 specific, which means strange things might happen. A private
4558 table is now used for this - though it costs 256 bytes, a table is
4559 much faster than multiple explicit tests. Of course, the standard
4560 character types table is still used for matching digits in subject
4561 strings against \d.
4563 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4564 ending in _t are). So I've renamed it as ESC_tee.
4566 3. The first argument for regexec() in the POSIX wrapper should have been
4567 defined as "const".
4569 4. Changed pcretest to use malloc() for its buffers so that they can be
4570 Electric Fenced for debugging.
4572 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4573 to read one or more bytes before the start of the subject string. Often this
4574 had no effect on PCRE's behaviour, but in some circumstances it could
4575 provoke a segmentation fault.
4577 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4578 to try to read one or more bytes before the start of the subject string.
4580 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4581 UTF-8 support could misbehave in various ways if the subject string
4582 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4583 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4584 back over UTF-8 characters.)
4587 Version 4.2 14-Apr-03
4588 ---------------------
4590 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4592 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4593 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4594 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4595 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4596 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4598 Note: automatic setting of the BUILD variables is not yet working
4599 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4600 compile-time but not at link-time
4601 [LINK]: use for linking executables only
4602 make different versions for Windows and non-Windows
4603 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4604 libraries
4605 [LINK_FOR_BUILD]: new variable
4606 [OBJEXT]: use throughout
4607 [EXEEXT]: use throughout
4608 <winshared>: new target
4609 <wininstall>: new target
4610 <dftables.o>: use native compiler
4611 <dftables>: use native linker
4612 <install>: handle Windows platform correctly
4613 <clean>: ditto
4614 <check>: ditto
4615 copy DLL to top builddir before testing
4617 As part of these changes, -no-undefined was removed again. This was reported
4618 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4619 in any case.
4621 3. Some tidies to get rid of compiler warnings:
4623 . In the match_data structure, match_limit was an unsigned long int, whereas
4624 match_call_count was an int. I've made them both unsigned long ints.
4626 . In pcretest the fact that a const uschar * doesn't automatically cast to
4627 a void * provoked a warning.
4629 . Turning on some more compiler warnings threw up some "shadow" variables
4630 and a few more missing casts.
4632 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4633 option, a class that contained a single character with a value between 128
4634 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4636 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4637 option, a class that contained several characters, but with at least one
4638 whose value was between 128 and 255 caused PCRE to crash.
4641 Version 4.1 12-Mar-03
4642 ---------------------
4644 1. Compiling with gcc -pedantic found a couple of places where casts were
4645 needed, and a string in dftables.c that was longer than standard compilers are
4646 required to support.
4648 2. Compiling with Sun's compiler found a few more places where the code could
4649 be tidied up in order to avoid warnings.
4651 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4652 first of these names is deprecated in the latest Autoconf in favour of the name
4653 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4654 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4655 analogy I have changed it to CFLAGS_FOR_BUILD.
4657 4. Added -no-undefined to the linking command in the Makefile, because this is
4658 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4659 linking step for the pcreposix library.
4661 5. PCRE was failing to diagnose the case of two named groups with the same
4662 name.
4664 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4665 literal character that is needed in the subject for a match, and scans along to
4666 ensure that it is present before embarking on the full matching process. This
4667 saves time in cases of nested unlimited repeats that are never going to match.
4668 Problem: the scan can take a lot of time if the subject is very long (e.g.
4669 megabytes), thus penalizing straightforward matches. It is now done only if the
4670 amount of subject to be scanned is less than 1000 bytes.
4672 7. A lesser problem with the same optimization is that it was recording the
4673 first character of an anchored pattern as "needed", thus provoking a search
4674 right along the subject, even when the first match of the pattern was going to
4675 fail. The "needed" character is now not set for anchored patterns, unless it
4676 follows something in the pattern that is of non-fixed length. Thus, it still
4677 fulfils its original purpose of finding quick non-matches in cases of nested
4678 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4681 Version 4.0 17-Feb-03
4682 ---------------------
4684 1. If a comment in an extended regex that started immediately after a meta-item
4685 extended to the end of string, PCRE compiled incorrect data. This could lead to
4686 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4688 2. Moved to autoconf 2.53 and libtool 1.4.2.
4690 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4691 the special perltest8 script is no longer needed - all the tests can be run
4692 from a single perltest script.
4694 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4695 by \s. It has now been removed in PCRE. This means it isn't recognized as
4696 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4697 class [:space:] *does* include VT, thereby creating a mess.
4699 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4700 space and tab.
4702 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4703 its new features into the main test script, reducing the number of scripts.
4705 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4706 were backward compatible, and made the (?i) apply to the whole pattern, as if
4707 /i were given. Now it behaves more logically, and applies the option setting
4708 only to what follows. PCRE has been changed to follow suit. However, if it
4709 finds options settings right at the start of the pattern, it extracts them into
4710 the global options, as before. Thus, they show up in the info data.
4712 8. Added support for the \Q...\E escape sequence. Characters in between are
4713 treated as literals. This is slightly different from Perl in that $ and @ are
4714 also handled as literals inside the quotes. In Perl, they will cause variable
4715 interpolation. Note the following examples:
4717 Pattern PCRE matches Perl matches
4719 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4720 \Qabc\$xyz\E abc\$xyz abc\$xyz
4721 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4723 For compatibility with Perl, \Q...\E sequences are recognized inside character
4724 classes as well as outside them.
4726 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4727 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4728 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4729 signed/unsigned warnings.
4731 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4732 option for pcretest, so I've replaced it by a simple function that does just
4733 that job.
4735 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4736 "pcregrep -".
4738 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4739 Java package. This provides some syntactic sugar for simple cases of what my
4740 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4741 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4742 item, you can use this simplified notation. Note that only makes sense with
4743 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4744 greediness, whatever the setting of the PCRE_UNGREEDY option.
4746 13. A change of greediness default within a pattern was not taking effect at
4747 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4748 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4749 was abstracted outside.
4751 14. PCRE now supports the \G assertion. It is true when the current matching
4752 position is at the start point of the match. This differs from \A when the
4753 starting offset is non-zero. Used with the /g option of pcretest (or similar
4754 code), it works in the same way as it does for Perl's /g option. If all
4755 alternatives of a regex begin with \G, the expression is anchored to the start
4756 match position, and the "anchored" flag is set in the compiled expression.
4758 15. Some bugs concerning the handling of certain option changes within patterns
4759 have been fixed. These applied to options other than (?ims). For example,
4760 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4761 been the other way round. Some of this was related to change 7 above.
4763 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4764 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4765 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4766 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4768 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4769 mode. Unlike ".", it always matches newline, whatever the setting of
4770 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4771 assertions. Perl allows it, but it doesn't (in general) work because it can't
4772 calculate the length of the lookbehind. At least, that's the case for Perl
4773 5.8.0 - I've been told they are going to document that it doesn't work in
4774 future.
4776 18. Added an error diagnosis for escapes that PCRE does not support: these are
4777 \L, \l, \N, \P, \p, \U, \u, and \X.
4779 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4780 reading past the end of the pattern in cases such as /[abcd/.
4782 20. PCRE was getting more memory than necessary for patterns with classes that
4783 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4785 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4786 compiling PCRE for use with Virtual Pascal.
4788 22. Small fix to the Makefile to make it work properly if the build is done
4789 outside the source tree.
4791 23. Added a new extension: a condition to go with recursion. If a conditional
4792 subpattern starts with (?(R) the "true" branch is used if recursion has
4793 happened, whereas the "false" branch is used only at the top level.
4795 24. When there was a very long string of literal characters (over 255 bytes
4796 without UTF support, over 250 bytes with UTF support), the computation of how
4797 much memory was required could be incorrect, leading to segfaults or other
4798 strange effects.
4800 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4801 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4802 there was a subsequent back reference to those brackets. This meant that, for
4803 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4804 possible to check for precisely this case. All we can do is abandon the
4805 optimization if .* occurs inside capturing brackets when there are any back
4806 references whatsoever. (See below for a better fix that came later.)
4808 26. The handling of the optimization for finding the first character of a
4809 non-anchored pattern, and for finding a character that is required later in the
4810 match were failing in some cases. This didn't break the matching; it just
4811 failed to optimize when it could. The way this is done has been re-implemented.
4813 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4815 28. Added a new feature that provides some of the functionality that Perl
4816 provides with (?{...}). The facility is termed a "callout". The way it is done
4817 in PCRE is for the caller to provide an optional function, by setting
4818 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4819 global variable. By default it is unset, which disables all calling out. To get
4820 the function called, the regex must include (?C) at appropriate points. This
4821 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4822 This provides a means of identifying different callout points. When PCRE
4823 reaches such a point in the regex, if pcre_callout has been set, the external
4824 function is called. It is provided with data in a structure called
4825 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4826 matching continues; if it returns a non-zero value, the match at the current
4827 point fails. However, backtracking will occur if possible. [This was changed
4828 later and other features added - see item 49 below.]
4830 29. pcretest is upgraded to test the callout functionality. It provides a
4831 callout function that displays information. By default, it shows the start of
4832 the match and the current position in the text. There are some new data escapes
4833 to vary what happens:
4835 \C+ in addition, show current contents of captured substrings
4836 \C- do not supply a callout function
4837 \C!n return 1 when callout number n is reached
4838 \C!n!m return 1 when callout number n is reached for the mth time
4840 30. If pcregrep was called with the -l option and just a single file name, it
4841 output "<stdin>" if a match was found, instead of the file name.
4843 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4844 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4845 pcre_exec(). This saves a malloc/free per call. The default value of
4846 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4847 when configuring.
4849 32. The default maximum size of a compiled pattern is 64K. There have been a
4850 few cases of people hitting this limit. The code now uses macros to handle the
4851 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4852 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4853 configuring. Tests 2 and 5 work only with 2-byte links because they output
4854 debugging information about compiled patterns.
4856 33. Internal code re-arrangements:
4858 (a) Moved the debugging function for printing out a compiled regex into
4859 its own source file (printint.c) and used #include to pull it into
4860 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4861 separate copies.
4863 (b) Defined the list of op-code names for debugging as a macro in
4864 internal.h so that it is next to the definition of the opcodes.
4866 (c) Defined a table of op-code lengths for simpler skipping along compiled
4867 code. This is again a macro in internal.h so that it is next to the
4868 definition of the opcodes.
4870 34. Added support for recursive calls to individual subpatterns, along the
4871 lines of Robin Houston's patch (but implemented somewhat differently).
4873 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4874 allow it to read and process whole directories in Win32. This code was
4875 contributed by Lionel Fourquaux; it has not been tested by me.
4877 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4878 used to name a group. Names consist of alphanumerics and underscores, and must
4879 be unique. Back references use the syntax (?P=name) and recursive calls use
4880 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4881 numbers. The function pcre_fullinfo() can be used after compilation to extract
4882 a name/number map. There are three relevant calls:
4884 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4885 PCRE_INFO_NAMECOUNT yields the number of entries
4886 PCRE_INFO_NAMETABLE yields a pointer to the map.
4888 The map is a vector of fixed-size entries. The size of each entry depends on
4889 the length of the longest name used. The first two bytes of each entry are the
4890 group number, most significant byte first. There follows the corresponding
4891 name, zero terminated. The names are in alphabetical order.
4893 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4894 case instead of 255. Making it the same both with and without UTF-8 support
4895 means that the same test output works with both.
4897 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4898 calling malloc() with a zero argument.
4900 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4901 optimization. I've improved things by keeping a bitmap of backreferences with
4902 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4903 fact referenced, the optimization can be applied. It is unlikely that a
4904 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4905 the match to follow \n) will appear inside brackets with a number greater than
4906 31, but if it does, any back reference > 31 suppresses the optimization.
4908 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4909 of disabling numbered capturing parentheses. Any opening parenthesis that is
4910 not followed by ? behaves as if it were followed by ?: but named parentheses
4911 can still be used for capturing (and they will acquire numbers in the usual
4912 way).
4914 41. Redesigned the return codes from the match() function into yes/no/error so
4915 that errors can be passed back from deep inside the nested calls. A malloc
4916 failure while inside a recursive subpattern call now causes the
4917 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4919 42. It is now possible to set a limit on the number of times the match()
4920 function is called in a call to pcre_exec(). This facility makes it possible to
4921 limit the amount of recursion and backtracking, though not in a directly
4922 obvious way, because the match() function is used in a number of different
4923 circumstances. The count starts from zero for each position in the subject
4924 string (for non-anchored patterns). The default limit is, for compatibility, a
4925 large number, namely 10 000 000. You can change this in two ways:
4927 (a) When configuring PCRE before making, you can use --with-match-limit=n
4928 to set a default value for the compiled library.
4930 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4931 a different value is set. See 45 below.
4933 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4935 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4936 of things that can be changed at compile time. The first argument specifies
4937 what is wanted and the second points to where the information is to be placed.
4938 The current list of available information is:
4942 The output is an integer that is set to one if UTF-8 support is available;
4943 otherwise it is set to zero.
4947 The output is an integer that it set to the value of the code that is used for
4948 newline. It is either LF (10) or CR (13).
4952 The output is an integer that contains the number of bytes used for internal
4953 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4957 The output is an integer that contains the threshold above which the POSIX
4958 interface uses malloc() for output vectors. See item 31 above.
4962 The output is an unsigned integer that contains the default limit of the number
4963 of match() calls in a pcre_exec() execution. See 42 above.
4965 44. pcretest has been upgraded by the addition of the -C option. This causes it
4966 to extract all the available output from the new pcre_config() function, and to
4967 output it. The program then exits immediately.
4969 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4970 order to support additional features. One way would have been to define
4971 pcre_exec2() (for example) with extra arguments, but this would not have been
4972 extensible, and would also have required all calls to the original function to
4973 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4974 is used for passing in "extra" data from pcre_study().
4976 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4977 contains the following fields:
4979 flags a bitmap indicating which of the following fields are set
4980 study_data opaque data from pcre_study()
4981 match_limit a way of specifying a limit on match() calls for a specific
4982 call to pcre_exec()
4983 callout_data data for callouts (see 49 below)
4985 The flag bits are also defined in pcre.h, and are
4991 The pcre_study() function now returns one of these new pcre_extra blocks, with
4992 the actual study data pointed to by the study_data field, and the
4993 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4994 before. That is, this change is entirely upwards-compatible and requires no
4995 change to existing code.
4997 If you want to pass in additional data to pcre_exec(), you can either place it
4998 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4999 block.
5001 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5002 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5003 times with different match limits, until it finds the minimum value needed for
5004 pcre_exec() to complete. The value is then output. This can be instructive; for
5005 most simple matches the number is quite small, but for pathological cases it
5006 gets very large very quickly.
5008 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5009 returns the size of the data block pointed to by the study_data field in a
5010 pcre_extra block, that is, the value that was passed as the argument to
5011 pcre_malloc() when PCRE was getting memory in which to place the information
5012 created by pcre_study(). The fourth argument should point to a size_t variable.
5013 pcretest has been extended so that this information is shown after a successful
5014 pcre_study() call when information about the compiled regex is being displayed.
5016 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5017 because what follows is always an absolute path. (Later: it turns out that this
5018 is more than cosmetic for MinGW, because it doesn't like empty path
5019 components.)
5021 49. Some changes have been made to the callout feature (see 28 above):
5023 (i) A callout function now has three choices for what it returns:
5025 0 => success, carry on matching
5026 > 0 => failure at this point, but backtrack if possible
5027 < 0 => serious error, return this value from pcre_exec()
5029 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5030 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5031 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5032 use by callout functions. It will never be used by PCRE itself.
5034 (ii) The pcre_extra structure (see 45 above) has a void * field called
5035 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5036 pcre_callout_block structure has a field of the same name. The contents of
5037 the field passed in the pcre_extra structure are passed to the callout
5038 function in the corresponding field in the callout block. This makes it
5039 easier to use the same callout-containing regex from multiple threads. For
5040 testing, the pcretest program has a new data escape
5042 \C*n pass the number n (may be negative) as callout_data
5044 If the callout function in pcretest receives a non-zero value as
5045 callout_data, it returns that value.
5047 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5048 there were some redundant $(CFLAGS) in commands that are now specified as
5049 $(LINK), which already includes $(CFLAGS).
5051 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5052 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5053 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5054 one-byte characters throughout. Note that case-insensitive matching applies
5055 only to characters whose values are less than 256. PCRE doesn't support the
5056 notion of cases for higher-valued characters.
5058 (i) A character class whose characters are all within 0-255 is handled as
5059 a bit map, and the map is inverted for negative classes. Previously, a
5060 character > 255 always failed to match such a class; however it should
5061 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5063 (ii) A negated character class with a single character < 255 is coded as
5064 "not this character" (OP_NOT). This wasn't working properly when the test
5065 character was multibyte, either singly or repeated.
5067 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5068 mode, for example: \x{100}{2,3}.
5070 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5071 singly or repeated) now correctly test multibyte characters. However,
5072 PCRE doesn't recognize any characters with values greater than 255 as
5073 digits, spaces, or word characters. Such characters always match \D, \S,
5074 and \W, and never match \d, \s, or \w.
5076 (v) Classes may now contain characters and character ranges with values
5077 greater than 255. For example: [ab\x{100}-\x{400}].
5079 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5080 PCRE in UTF-8 mode.
5082 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5083 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5084 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5085 value.)
5087 53. The single man page has become too large. I have therefore split it up into
5088 a number of separate man pages. These also give rise to individual HTML pages;
5089 these are now put in a separate directory, and there is an index.html page that
5090 lists them all. Some hyperlinking between the pages has been installed.
5092 54. Added convenience functions for handling named capturing parentheses.
5094 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5095 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5096 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5097 are faulted.
5099 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5100 calling configure. These values are used when compiling the dftables.c program
5101 which is run to generate the source of the default character tables. They
5102 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5103 you will need to set these values.
5105 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5108 Version 3.9 02-Jan-02
5109 ---------------------
5111 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5113 2. If --disable-static was given, the building process failed when trying to
5114 build pcretest and pcregrep. (For some reason it was using libtool to compile
5115 them, which is not right, as they aren't part of the library.)
5118 Version 3.8 18-Dec-01
5119 ---------------------
5121 1. The experimental UTF-8 code was completely screwed up. It was packing the
5122 bytes in the wrong order. How dumb can you get?
5125 Version 3.7 29-Oct-01
5126 ---------------------
5128 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5129 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5130 this didn't happen under Solaris 8, where I normally test things.
5132 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5133 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5134 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5135 this fix an hour or so after the initial 3.7 release.)
5138 Version 3.6 23-Oct-01
5139 ---------------------
5141 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5142 offsets passed as NULL with zero offset count.
5144 2. The config.guess and config.sub files had not been updated when I moved to
5145 the latest autoconf.
5148 Version 3.5 15-Aug-01
5149 ---------------------
5151 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5152 had been forgotten.
5154 2. By using declared but undefined structures, we can avoid using "void"
5155 definitions in pcre.h while keeping the internal definitions of the structures
5156 private.
5158 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5159 user point of view, this means that both static and shared libraries are built
5160 by default, but this can be individually controlled. More of the work of
5161 handling this static/shared cases is now inside libtool instead of PCRE's make
5162 file.
5164 4. The pcretest utility is now installed along with pcregrep because it is
5165 useful for users (to test regexs) and by doing this, it automatically gets
5166 relinked by libtool. The documentation has been turned into a man page, so
5167 there are now .1, .txt, and .html versions in /doc.
5169 5. Upgrades to pcregrep:
5170 (i) Added long-form option names like gnu grep.
5171 (ii) Added --help to list all options with an explanatory phrase.
5172 (iii) Added -r, --recursive to recurse into sub-directories.
5173 (iv) Added -f, --file to read patterns from a file.
5175 6. pcre_exec() was referring to its "code" argument before testing that
5176 argument for NULL (and giving an error if it was NULL).
5178 7. Upgraded Makefile.in to allow for compiling in a different directory from
5179 the source directory.
5181 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5182 options bits, the pointer it was passed was to an int instead of to an unsigned
5183 long int. This mattered only on 64-bit systems.
5185 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5186 generated) instead of pcre.in, which it its source. Also made the same change
5187 in several of the .c files.
5189 10. A new release of gcc defines printf() as a macro, which broke pcretest
5190 because it had an ifdef in the middle of a string argument for printf(). Fixed
5191 by using separate calls to printf().
5193 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5194 script, to force use of CR or LF instead of \n in the source. On non-Unix
5195 systems, the value can be set in config.h.
5197 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5198 absolute limit. Changed the text of the error message to make this clear, and
5199 likewise updated the man page.
5201 13. The limit of 99 on the number of capturing subpatterns has been removed.
5202 The new limit is 65535, which I hope will not be a "real" limit.
5205 Version 3.4 22-Aug-00
5206 ---------------------
5208 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5210 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5213 Version 3.3 01-Aug-00
5214 ---------------------
5216 1. If an octal character was given, but the value was greater than \377, it
5217 was not getting masked to the least significant bits, as documented. This could
5218 lead to crashes in some systems.
5220 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5221 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5223 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5224 These just pass their arguments on to (pcre_free)(), but they are provided
5225 because some uses of PCRE bind it to non-C systems that can call its functions,
5226 but cannot call free() or pcre_free() directly.
5228 4. Add "make test" as a synonym for "make check". Corrected some comments in
5229 the Makefile.
5231 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5232 Makefile.
5234 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5235 command called pgrep for grepping around the active processes.
5237 7. Added the beginnings of support for UTF-8 character strings.
5239 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5240 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5241 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5242 out for the ar command.)
5245 Version 3.2 12-May-00
5246 ---------------------
5248 This is purely a bug fixing release.
5250 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5251 of ZA. This was just one example of several cases that could provoke this bug,
5252 which was introduced by change 9 of version 2.00. The code for breaking
5253 infinite loops after an iteration that matches an empty string was't working
5254 correctly.
5256 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5257 when matched against abbab (for example). After matching an empty string, it
5258 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5259 caused it to match further down the string than it should.
5261 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5262 was there because it doesn't seem to be needed, and it causes trouble on some
5263 systems, as it is not a Standard C header. It has been removed.
5265 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5266 were reported on the Macintosh. The changes were from
5268 while ((c = *(++ptr)) != 0 && c != '\n');
5269 to
5270 while ((c = *(++ptr)) != 0 && c != '\n') ;
5272 Totally extraordinary, but if that's what it takes...
5274 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5275 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5276 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5277 assumes the way PCRE uses memmove() (always moving upwards).
5279 6. PCRE is being used in one environment where strchr() is not available. There
5280 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5281 faster code anyway.
5284 Version 3.1 09-Feb-00
5285 ---------------------
5287 The only change in this release is the fixing of some bugs in Makefile.in for
5288 the "install" target:
5290 (1) It was failing to install pcreposix.h.
5292 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5295 Version 3.0 01-Feb-00
5296 ---------------------
5298 1. Add support for the /+ modifier to perltest (to output $` like it does in
5299 pcretest).
5301 2. Add support for the /g modifier to perltest.
5303 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5304 matches null strings.
5306 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5307 pattern. Perl treats empty patterns specially - it reuses the most recent
5308 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5309 effect.
5311 5. The POSIX interface was broken in that it was just handing over the POSIX
5312 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5313 required a bigger vector, with some working space on the end. This means that
5314 the POSIX wrapper now has to get and free some memory, and copy the results.
5316 6. Added some simple autoconf support, placing the test data and the
5317 documentation in separate directories, re-organizing some of the
5318 information files, and making it build pcre-config (a GNU standard). Also added
5319 libtool support for building PCRE as a shared library, which is now the
5320 default.
5322 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5323 09 are not valid octal constants. Single digits will be used for minor values
5324 less than 10.
5326 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5327 existing programs that set these in the POSIX interface can use PCRE without
5328 modification.
5330 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5331 return all that pcre_info() returns, plus additional data. The pcre_info()
5332 function is retained for compatibility, but is considered to be obsolete.
5334 10. Added experimental recursion feature (?R) to handle one common case that
5335 Perl 5.6 will be able to do with (?p{...}).
5337 11. Added support for POSIX character classes like [:alpha:], which Perl is
5338 adopting.
5341 Version 2.08 31-Aug-99
5342 ----------------------
5344 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5345 trying to match at the startoffset position, but instead was moving forward to
5346 the next newline as if a previous match had failed.
5348 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5349 and could get into a loop if a null string was matched other than at the start
5350 of the subject.
5352 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5353 be distinguished at compile time, and for completeness also added PCRE_DATE.
5355 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5356 in GnuWin32 environments.
5359 Version 2.07 29-Jul-99
5360 ----------------------
5362 1. The documentation is now supplied in plain text form and HTML as well as in
5363 the form of man page sources.
5365 2. C++ compilers don't like assigning (void *) values to other pointer types.
5366 In particular this affects malloc(). Although there is no problem in Standard
5367 C, I've put in casts to keep C++ compilers happy.
5369 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5370 should be (const char *).
5372 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5373 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5374 However, I haven't made this a standard facility. The documentation doesn't
5375 mention it, and the Makefile doesn't support it.
5377 5. The Makefile now contains an "install" target, with editable destinations at
5378 the top of the file. The pcretest program is not installed.
5380 6. pgrep -V now gives the PCRE version number and date.
5382 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5383 causing the entire string to be ignored, instead of just the last character.
5385 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5386 non-matching string, it can take a very, very long time, even for strings of
5387 quite modest length, because of the nested recursion. PCRE now does better in
5388 some of these cases. It does this by remembering the last required literal
5389 character in the pattern, and pre-searching the subject to ensure it is present
5390 before running the real match. In other words, it applies a heuristic to detect
5391 some types of certain failure quickly, and in the above example, if presented
5392 with a string that has no trailing " it gives "no match" very quickly.
5394 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5395 other alternatives are tried instead.
5398 Version 2.06 09-Jun-99
5399 ----------------------
5401 1. Change pcretest's output for amount of store used to show just the code
5402 space, because the remainder (the data block) varies in size between 32-bit and
5403 64-bit systems.
5405 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5406 start matching at. This allows lookbehinds to work when searching for multiple
5407 occurrences in a string.
5409 3. Added additional options to pcretest for testing multiple occurrences:
5411 /+ outputs the rest of the string that follows a match
5412 /g loops for multiple occurrences, using the new startoffset argument
5413 /G loops for multiple occurrences by passing an incremented pointer
5415 4. PCRE wasn't doing the "first character" optimization for patterns starting
5416 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5417 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5418 the letter 'x'. On long subject strings, this gives a significant speed-up.
5421 Version 2.05 21-Apr-99
5422 ----------------------
5424 1. Changed the type of magic_number from int to long int so that it works
5425 properly on 16-bit systems.
5427 2. Fixed a bug which caused patterns starting with .* not to work correctly
5428 when the subject string contained newline characters. PCRE was assuming
5429 anchoring for such patterns in all cases, which is not correct because .* will
5430 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5431 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5432 must be retried after every newline in the subject.
5435 Version 2.04 18-Feb-99
5436 ----------------------
5438 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5439 computation of the store needed to hold the pattern was incorrect (too large).
5440 If such patterns were nested a few deep, this could multiply and become a real
5441 problem.
5443 2. Added /M option to pcretest to show the memory requirement of a specific
5444 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5446 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5447 compiled in such a way that the backtracking after subsequent failure was
5448 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5449 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5452 Version 2.03 02-Feb-99
5453 ----------------------
5455 1. Fixed typo and small mistake in man page.
5457 2. Added 4th condition (GPL supersedes if conflict) and created separate
5458 LICENCE file containing the conditions.
5460 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5461 Perl, that is the internal \ allows the delimiter to be included in the
5462 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5463 the final delimiter, add \ to the end of the pattern (to test the error).
5465 4. Added the convenience functions for extracting substrings after a successful
5466 match. Updated pcretest to make it able to test these functions.
5469 Version 2.02 14-Jan-99
5470 ----------------------
5472 1. Initialized the working variables associated with each extraction so that
5473 their saving and restoring doesn't refer to uninitialized store.
5475 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5476 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5477 fix the problem.
5479 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5480 calls, and wasn't printing the correct value for compiling calls. Increased the
5481 default value of LOOPREPEAT, and the number of significant figures in the
5482 times.
5484 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5486 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5487 a building problem on Windows NT with a FAT file system.
5490 Version 2.01 21-Oct-98
5491 ----------------------
5493 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5494 to character tables built by pcre_maketables() in the current locale. If NULL
5495 is passed, the default tables are used.
5498 Version 2.00 24-Sep-98
5499 ----------------------
5501 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5502 it any more.
5504 2. Allow quantification of (?>) groups, and make it work correctly.
5506 3. The first character computation wasn't working for (?>) groups.
5508 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5509 end of the subject) and add 5.005's \z, which really does match only at the
5510 very end of the subject.
5512 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5514 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5515 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5516 localized options. All options to pcre_study() were also removed.
5518 7. Add other new features from 5.005:
5520 $(?<= positive lookbehind
5521 $(?<! negative lookbehind
5522 (?imsx-imsx) added the unsetting capability
5523 such a setting is global if at outer level; local otherwise
5524 (?imsx-imsx:) non-capturing groups with option setting
5525 (?(cond)re|re) conditional pattern matching
5527 A backreference to itself in a repeated group matches the previous
5528 captured string.
5530 8. General tidying up of studying (both automatic and via "study")
5531 consequential on the addition of new assertions.
5533 9. As in 5.005, unlimited repeated groups that could match an empty substring
5534 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5535 runtime if any iteration does actually match an empty substring.
5537 10. Include the RunTest script in the distribution.
5539 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5540 discrepancies, some of which were old and were also with respect to 5.004. They
5541 have now been fixed.
5544 Version 1.09 28-Apr-98
5545 ----------------------
5547 1. A negated single character class followed by a quantifier with a minimum
5548 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5549 program crashes, or just wrong answers. This did not apply to negated classes
5550 containing more than one character, or to minima other than one.
5553 Version 1.08 27-Mar-98
5554 ----------------------
5556 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5558 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5559 latter must appear before anything that relies on it in the pattern.
5562 Version 1.07 16-Feb-98
5563 ----------------------
5565 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5566 repeat of a potentially empty string).
5569 Version 1.06 23-Jan-98
5570 ----------------------
5572 1. Added Markus Oberhumer's little patches for C++.
5574 2. Literal strings longer than 255 characters were broken.
5577 Version 1.05 23-Dec-97
5578 ----------------------
5580 1. Negated character classes containing more than one character were failing if
5581 PCRE_CASELESS was set at run time.
5584 Version 1.04 19-Dec-97
5585 ----------------------
5587 1. Corrected the man page, where some "const" qualifiers had been omitted.
5589 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5590 input syntax.
5592 3. Fixed memory leak which occurred when a regex with back references was
5593 matched with an offsets vector that wasn't big enough. The temporary memory
5594 that is used in this case wasn't being freed if the match failed.
5596 4. Tidied pcretest to ensure it frees memory that it gets.
5598 5. Temporary memory was being obtained in the case where the passed offsets
5599 vector was exactly big enough.
5601 6. Corrected definition of offsetof() from change 5 below.
5603 7. I had screwed up change 6 below and broken the rules for the use of
5604 setjmp(). Now fixed.
5607 Version 1.03 18-Dec-97
5608 ----------------------
5610 1. A erroneous regex with a missing opening parenthesis was correctly
5611 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5612 on some systems.
5614 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5615 it was reported that one broken compiler failed on the former because "code" is
5616 also an independent variable.
5618 3. The erroneous regex a[]b caused an array overrun reference.
5620 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5621 fail on data ending with that character. (It was going on too far, and checking
5622 the next character, typically a binary zero.) This was specific to the
5623 optimized code for single-character negative classes.
5625 5. Added a contributed patch from the TIN world which does the following:
5627 + Add an undef for memmove, in case the the system defines a macro for it.
5629 + Add a definition of offsetof(), in case there isn't one. (I don't know
5630 the reason behind this - offsetof() is part of the ANSI standard - but
5631 it does no harm).
5633 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5634 most of the places where whitespace preceded '#'. I have given up and
5635 allowed the remaining 2 cases to be at the margin.
5637 + Rename some variables in pcre to eliminate shadowing. This seems very
5638 pedantic, but does no harm, of course.
5640 6. Moved the call to setjmp() into its own function, to get rid of warnings
5641 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5643 7. Constructs such as \d{8,} were compiling into the equivalent of
5644 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5645 outcome, but in this particular case used more store than had been allocated,
5646 which caused the bug to be discovered because it threw up an internal error.
5648 8. The debugging code in both pcre and pcretest for outputting the compiled
5649 form of a regex was going wrong in the case of back references followed by
5650 curly-bracketed repeats.
5653 Version 1.02 12-Dec-97
5654 ----------------------
5656 1. Typos in pcre.3 and comments in the source fixed.
5658 2. Applied a contributed patch to get rid of places where it used to remove
5659 'const' from variables, and fixed some signed/unsigned and uninitialized
5660 variable warnings.
5662 3. Added the "runtest" target to Makefile.
5664 4. Set default compiler flag to -O2 rather than just -O.
5667 Version 1.01 19-Nov-97
5668 ----------------------
5670 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5671 like /([ab]*)*/, that is, for classes with more than one character in them.
5673 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5674 as /((?>a*))*/ (a PCRE_EXTRA facility).
5677 Version 1.00 18-Nov-97
5678 ----------------------
5680 1. Added compile-time macros to support systems such as SunOS4 which don't have
5681 memmove() or strerror() but have other things that can be used instead.
5683 2. Arranged that "make clean" removes the executables.
5686 Version 0.99 27-Oct-97
5687 ----------------------
5689 1. Fixed bug in code for optimizing classes with only one character. It was
5690 initializing a 32-byte map regardless, which could cause it to run off the end
5691 of the memory it had got.
5693 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5696 Version 0.98 22-Oct-97
5697 ----------------------
5699 1. Fixed bug in code for handling temporary memory usage when there are more
5700 back references than supplied space in the ovector. This could cause segfaults.
5703 Version 0.97 21-Oct-97
5704 ----------------------
5706 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5708 2. Optimized negated single characters not to use a bit map.
5710 3. Brought error texts together as macro definitions; clarified some of them;
5711 fixed one that was wrong - it said "range out of order" when it meant "invalid
5712 escape sequence".
5714 4. Changed some char * arguments to const char *.
5716 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5718 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5719 pcretest.
5722 Version 0.96 16-Oct-97
5723 ----------------------
5725 1. Added a simple "pgrep" utility to the distribution.
5727 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5728 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5729 where "ddd" means "one or more decimal digits".
5731 3. Fixed serious bug. If a pattern had a back reference, but the call to
5732 pcre_exec() didn't supply a large enough ovector to record the related
5733 identifying subpattern, the match always failed. PCRE now remembers the number
5734 of the largest back reference, and gets some temporary memory in which to save
5735 the offsets during matching if necessary, in order to ensure that
5736 backreferences always work.
5738 4. Increased the compatibility with Perl in a number of ways:
5740 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5741 to request this handling. The option can be set at compile or exec time.
5743 (b) $ matches before a terminating newline by default; an option
5744 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5745 mode). The option can be set at compile or exec time.
5747 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5748 the same as Perl's. If the decimal number it represents is less than 10
5749 or there aren't that many previous left capturing parentheses, an octal
5750 escape is read. Inside a character class, it's always an octal escape,
5751 even if it is a single digit.
5753 (d) An escaped but undefined alphabetic character is taken as a literal,
5754 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5755 escapes.
5757 (e) {0} is now permitted. (The previous item is removed from the compiled
5758 pattern).
5760 5. Changed all the names of code files so that the basic parts are no longer
5761 than 10 characters, and abolished the teeny "globals.c" file.
5763 6. Changed the handling of character classes; they are now done with a 32-byte
5764 bit map always.
5766 7. Added the -d and /D options to pcretest to make it possible to look at the
5767 internals of compilation without having to recompile pcre.
5770 Version 0.95 23-Sep-97
5771 ----------------------
5773 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5774 \x20 at the start of a run of normal characters. These were being treated as
5775 real characters, instead of the source characters being re-checked.
5778 Version 0.94 18-Sep-97
5779 ----------------------
5781 1. The functions are now thread-safe, with the caveat that the global variables
5782 containing pointers to malloc() and free() or alternative functions are the
5783 same for all threads.
5785 2. Get pcre_study() to generate a bitmap of initial characters for non-
5786 anchored patterns when this is possible, and use it if passed to pcre_exec().
5789 Version 0.93 15-Sep-97
5790 ----------------------
5792 1. /(b)|(:+)/ was computing an incorrect first character.
5794 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5795 but not actually doing anything yet.
5797 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5798 as Perl does (e.g. [-az] or [az-]).
5800 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5801 all possible positions.
5803 5. Split up into different modules to avoid including unneeded functions in a
5804 compiled binary. However, compile and exec are still in one module. The "study"
5805 function is split off.
5807 6. The character tables are now in a separate module whose source is generated
5808 by an auxiliary program - but can then be edited by hand if required. There are
5809 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5810 toupper() in the code.
5812 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5813 make them global. Abolish the function for setting them, as the caller can now
5814 set them directly.
5817 Version 0.92 11-Sep-97
5818 ----------------------
5820 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5821 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5823 2. Caseless matching was not working in character classes if the characters in
5824 the pattern were in upper case.
5826 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5828 4. Make PCRE_ANCHORED public and accept as a compile option.
5830 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5831 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5832 pass them.
5834 6. Give an error if bad option bits passed at compile or run time.
5836 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5837 pcretest to cause it to pass that flag.
5839 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5840 options, and the first character, if set.
5842 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5845 Version 0.91 10-Sep-97
5846 ----------------------
5848 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5849 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5851 2. PCRE was looping on encountering an indefinitely repeated back reference to
5852 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5853 Perl does - treats the match as successful.
5855 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5